Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
Neuroprotective Drug Delivery to the Injured Spinal Cord with Hyaluronan and Methylcellulose
by
Catherine Elizabeth Kang
A thesis submitted in conformity with the requirements for the degree of Doctorate of Philosophy
Department of Chemical Engineering and Applied Chemistry in collaboration with the Institute for Biomaterials and Biomedical Engineering
University of Toronto
© Copyright by Catherine Elizabeth Kang 2010
ii
Neuroprotective Drug Delivery to the Injured Spinal Cord with
Hyaluronan and Methylcellulose
Catherine Elizabeth Kang
Doctorate of Philosophy
Department of Chemical Engineering and Applied Chemistry in Collaboration with the Institute of Biomaterials and Biomedical Engineering
University of Toronto
2010
Abstract
Traumatic spinal cord injury (SCI) is a devastating condition for which there is no effective
clinical treatment. Neuroprotective molecules that minimize tissue loss have shown promising
results; however systemic delivery may limit in vivo benefits due to short systemic half-life and
minimal passage across the blood-spinal cord barrier. To overcome these limitations, an
injectable intrathecal delivery vehicle comprised of hyaluronan and methylcellulose (HAMC)
was developed, and previously demonstrated to be safe and biocompatible intrathecally. Here,
HAMC was determined to persist in the intrathecal space for between 4-7 d in vivo, indicating it
as an optimal delivery system for neuroprotective agents to reduce tissue degeneration after SCI.
HAMC was then investigated as an in vivo delivery system for two neuroprotective proteins:
erythropoietin (EPO) and fibroblast growth factor 2 (FGF2). Both proteins demonstrated a
diffusive release profile in vitro and maintained significant bioactivity during release. When
EPO was delivered intrathecally with HAMC to the injured spinal cord, reduced cavitation in the
tissue and significantly improved neuron counts were observed relative to the conventional
delivery strategies of intraperitoneal and intrathecal bolus. When FGF2 was delivered
iii
intrathecally from HAMC, therapeutic concentrations penetrated into the injured spinal cord
tissue for up to 6 h. Poly(ethylene glycol) modification of FGF2 significantly increased the
amount of protein that diffused into the tissue when delivered similarly. Because FGF2 is a
known angiogenic agent, dynamic computed tomography was developed for small animal serial
assessment of spinal cord hemodynamics. Following SCI and treatment with FGF2 from
HAMC, moderate improvement of spinal cord blood flow and a reduction in permeability were
observed up to 7 d post-injury, suggesting that early delivery of neuroprotective agents can have
lasting effects on tissue recovery. Importantly, the entirety of this work demonstrates that
HAMC is an effective short-term delivery system for neuroprotective agents by improving tissue
outcomes following traumatic SCI.
iv
Acknowledgments Science is an endeavor in which a single person cannot take complete credit for their work; as scientists and engineers, we stand on the shoulders of scientists and engineers who came before us to pave our way. As graduate students, we stand on the shoulders of those who support and guide us through the many years of diligent hard work. To each of you on whose shoulders I have stood, thank you.
I would like to thank Dr. Molly Shoichet for giving me the opportunity to work with her diverse team of engineers, neurosurgeons and chemists in the lab and for having confidence in me through the last several years. Your collaborative efforts truly shaped and provided breadth to my thesis. Thank you to Dr. Charles Tator for encouraging me to tackle difficult problems and providing invaluable guidance throughout my research. I am grateful to Dr. Michael Sefton for not being afraid to challenge me at each committee meeting, and motivating me to improve my depth of knowledge. I appreciate the patience of Dr. Ivan Yeung for allowing me to explore a new technique in a field outside my expertise. I would also like to thank Dr. Grant Allen for your role in my oral examinations, and for providing extracurricular opportunities to further my professional and academic goals. Finally, thank you to Dr. Ravi Bellamkonda for providing an honest evaluation of my thesis work. I am indebted to all of you who have contributed to my experience and my research here at the University of Toronto.
My experience here has been shaped through research with Team Injectable, fun times in the Cluster, and fruitful discussion with the many members of SBERT. More than anyone in Toronto, I want to thank Karyn Ho for her friendship and moral support when I needed it most – I can never pick up my guitar without thinking of you! I am genuinely appreciative to those who went above and beyond to make my time in Toronto more than enjoyable – Douglas Baumann, Jennifer Hudgins, Tina Bhinder, Ryan Wylie, Howard Kim, Yukie Aizawa, and Michael Conrad, and Peter Poon. I could have never called Toronto home without each of you. Also, thank you to those who have made my time here both profitable and productive – Kat Vulic, Ankeeta Tadkase, Yuanfei Wang, Valerie Peng, and Siang Lee. Your dedication to research was a source of motivation for me.
Finally, my deepest gratitude goes to the one who stayed with me through immeasurable heartache, all the while providing immense support – my husband, Peter Kang. I may never be able to adequately express the profound love and appreciation that I have for you. I never could have made it this far without you and I look forward to the days we will spend together as a family with our little OranguKang. Thanks baby Aravis for letting mommy finish her Ph.D. in the first days of your life! Thank you to my parents Jeff and Cathy Clark and to Mei Wan Kang for helping Peter and I in every way you could. Finally, thank you to Veronica Veale for being such a great friend and always knowing what to say to make me feel better.
Catherine E. Kang
v
Table of Contents
Acknowledgments .......................................................................................................................... iv
Table of Contents .............................................................................................................................v
List of Tables ................................................................................................................................. ix
List of Figures ..................................................................................................................................x
List of Appendices ....................................................................................................................... xiv
1 Background & Introduction ........................................................................................................1
1.1 Rationale ..............................................................................................................................1
1.2 Hypothesis and Objectives ...................................................................................................2
1.3 Anatomy and Physiology of the Human Spinal Cord ..........................................................3
1.4 Pathology and Prevalence of Spinal Cord Injury .................................................................5
1.5 Animal Models of Spinal Cord Injury .................................................................................8
1.5.1 Measurement Techniques for Spinal Cord Hemodynamics ..................................10
1.5.2 Motor Function Outcome Measures ......................................................................12
1.6 Treatment Strategies for Spinal Cord Injury Repair ..........................................................14
1.6.1 Clinical Trials and Treatments ...............................................................................14
1.6.2 Experimental Neuroprotective Agents ...................................................................16
1.6.3 Routes of Drug Administration ..............................................................................20
1.7 Polymers for Drug Delivery ...............................................................................................22
1.7.1 Hyaluronan .............................................................................................................24
1.7.2 Methylcellulose ......................................................................................................25
1.7.3 Poly(ethylene glycol) .............................................................................................27
1.8 Historical Context and Motivation for this Research ........................................................28
2 Localized Delivery of Erythropoietin with the Hyaluronan-Methylcellulose Blend ................31
vi
2.1 Abstract (Local Delivery of EPO) .....................................................................................31
2.2 Introduction (Local Delivery of EPO) ...............................................................................32
2.3 Materials & Methods (Local Delivery of EPO) .................................................................34
2.3.1 Material preparation ...............................................................................................34
2.3.2 In vitro release of EPO from HAMC .....................................................................34
2.3.3 Bioactivity of released EPO ...................................................................................35
2.3.4 Degradation of HAMC in vitro and in vivo ...........................................................35
2.3.5 EPO Efficacy Study ...............................................................................................37
2.3.6 Functional Assessment ...........................................................................................38
2.3.7 Histology & Immunohistochemistry ......................................................................38
2.3.8 Statistical Analysis .................................................................................................39
2.4 Results (Local Delivery of EPO) .......................................................................................39
2.4.1 In vitro release of EPO from HAMC .....................................................................39
2.4.2 Degradation of HAMC in vitro and in vivo ...........................................................41
2.4.3 In vivo delivery of EPO with HAMC ....................................................................43
2.5 Discussion (Local Delivery of EPO) .................................................................................47
3 Enhancing Localized Delivery of Fibroblast Growth Factor 2 .................................................51
3.1 Abstract (Penetration of FGF2) .........................................................................................51
3.2 Introduction (Penetration of FGF2) ...................................................................................51
3.3 Materials & Methods (Penetration of FGF2) .....................................................................53
3.3.1 Synthesis of PEG-FGF2 .........................................................................................53
3.3.2 Bioactivity of PEGylated FGF2 .............................................................................54
3.3.3 In vitro release of FGF2 vs. PEG-FGF2 from HAMC ..........................................55
3.3.4 In vivo distribution of FGF2 in spinal cord tissue ..................................................55
3.3.5 Statistical Analysis .................................................................................................58
3.4 Results (Penetration of FGF2) ...........................................................................................58
vii
3.4.1 PEGylation of FGF2 ..............................................................................................58
3.4.2 PEG-FGF2 Bioactivity ...........................................................................................60
3.4.3 FGF2 Release from HAMC In vitro ......................................................................60
3.4.4 In vivo Distribution of FGF2 and PEG-FGF2 ........................................................61
3.5 Discussion (Penetration of FGF2) .....................................................................................64
3.6 Conclusions (Penetration of FGF2) ...................................................................................66
4 Live Animal Imaging of Post-Injury Spinal Cord Blood Flow ................................................67
4.1 Abstract (Hemodynamics with FGF2) ...............................................................................67
4.2 Introduction (Hemodynamics with FGF2) .........................................................................68
4.3 Materials & Methods (Hemodynamics with FGF2) ..........................................................72
4.3.1 HAMC preparation ................................................................................................73
4.3.2 In vivo Surgical Procedures ...................................................................................73
4.3.3 Computed Tomography Imaging Acquisition with Dynamic Contrast Enhancement ..........................................................................................................74
4.3.4 Image Analysis .......................................................................................................74
4.3.5 Statistical analysis ..................................................................................................75
4.4 Results (Hemodynamics with FGF2) .................................................................................77
4.4.1 SCBF and PS Measurements in the Uninjured Spinal Cord .................................77
4.4.2 Epicenter of Injury (T1/T2) ...................................................................................77
4.4.3 0.75-1 mm Rostral to Injury Epicenter (T1) ..........................................................80
4.4.4 1.8-2 mm Rostral to Injury Epicenter (T1/C7) ......................................................82
4.5 Discussion (Hemodynamics with FGF2) ...........................................................................84
5 Discussion .................................................................................................................................87
5.1 Localized Drug Delivery ....................................................................................................87
5.1.1 In vitro Release of Therapeutic Molecules from HAMC ......................................88
5.1.2 In vivo Delivery from HAMC ................................................................................91
viii
5.2 Neuroprotection Achieved with Localized Delivery .........................................................96
5.2.1 Implications of EPO Crossing the BSCB ..............................................................97
5.2.2 Improvements in SCBF and Permeability with FGF2 Delivery ............................98
5.3 PEGylation for Improving Tissue Delivery .....................................................................100
5.4 Spinal Cord Blood Flow Imaging Strategies ...................................................................101
5.5 Acheivement of Objectives ..............................................................................................103
5.6 Conclusions ......................................................................................................................105
6 Limitations and Recommendations for Future Work .............................................................106
6.1 Extending Delivery Window for HAMC .........................................................................106
6.1.1 Define Mechanism for HAMC Elimination from CNS .......................................106
6.1.2 Extend In vivo Lifetime of HAMC ......................................................................107
6.1.3 Controlled Release ...............................................................................................107
6.2 Drug Choice Matrix .........................................................................................................108
6.3 Development of CT for SC hemodynamics .....................................................................109
6.4 Limitations for Bench to Bedside Translation .................................................................110
References & Bibliography ..........................................................................................................112
Appendix 1: Glossary ..................................................................................................................137
Appendix 2: Abbreviations ..........................................................................................................139
Appendix 3: Additional Data .......................................................................................................142
ix
List of Tables Table 1.1: Secondary injury timeline .............................................................................................. 6
Table 1.2: Basso, Beattie and Bresnahan (BBB) locomotor score rankings. ............................... 13
Table 1.3: Drugs for SCI treatment in cinical or experimental studies indicating potential benefits
and limitations ............................................................................................................................... 17
x
List of Figures Figure 1.1: Cross section of spinal cord showing the multiple layers of protection in CNS
(Copyright Michael Corrin). ........................................................................................................... 4
Figure 1.2: Illustration of leading models of spinal cord injury (Copyright Michael Corrin). ..... 10
Figure 1.3: Illustrations of drug delivery routes: (A) Intravenous [201], (B) intrathecal bolus via
lumbar puncture [202], and (C) implanted minipump (copyright Michael Corrin) ..................... 22
Figure 1.4: Chemical structure of hyaluronan (A), methylcellulose (B), and poly(ethylene glycol)
(C) ................................................................................................................................................. 25
Figure 1.5: Gelation of MC caused by heating via formation of hydrophobic junctions ............. 26
Figure 1.6: Injectable polymeric drug delivery system for intrathecal delivery (copyright Micheal
Corrin). .......................................................................................................................................... 30
Figure 2.1: (A) EPO release from HAMC over 128 h shows fast release within 16 h. (B)
Cumulative release of EPO vs. the square root of time demonstrates a diffusion-mediated release
profile for the first 75% of EPO released. (C) EPO released from HAMC is 80% bioactive over
time relative to cumulative EPO released. (=Cumulative EPO released; =Bioactive portion
of EPO released) ........................................................................................................................... 40
Figure 2.2 (see next page): (A) In vitro mass degradation is the same for labeled and unlabeled
HAMC. (B) Brightfield image of a parasagittal spinal cord section with corresponding
fluorescent images of HA (Green) and MC (Red) in the IT space immediately after injection.
(C) Loss in the area of fluorescence shows that HA degrades faster than MC in the IT space.
Scale bars =1mm. (n=3, mean ± standard deviation plotted) ((=mass loss of labeled HAMC;
=mass loss of unlabeled HAMC) =fluorescence loss of HA; =fluorescence loss of MC) 41
Figure 2.3: (A) Parasaggital section of a representative spinal cord of an animal that received IT
administration of EPO-loaded HAMC at maximal cavitation stained with LFB/H&E. (B)
Comparison of maximal cavity area measured from parasaggital spinal cord sections. (n=4,
xi
mean + standard deviation; horizontal bars show statistical significant based on a t-test, *p<0.05)
....................................................................................................................................................... 44
Figure 2.4: Neuron counts observed in the ventral horns caudal to the injury site based on NeuN
staining. (n=3, mean + standard deviation; horizontal bars show statistical significance based on
t=test, *p<0.05) ............................................................................................................................. 45
Figure 2.5: Comparison of maximal inflammation area measured with ED-1 immunostaining of
macrophages and microglia from parasaggital spinal cord sections. (n=3, mean ± standard
deviation). ..................................................................................................................................... 46
Figure 2.6: Open field motor scores as assessed by the BBB scale over six weeks. (*shows
statistical significance measured by ANOVA; n=9 to 11, data shown as mean ± standard
deviation) (=IT administration of EPO-loaded HAMC; =IT administration of HAMC alone;
=IT administration of EPO; =IP administration of EPO) ...................................................... 47
Figure 3.1: (A) Diagram of injection paradigm shows hydrogel spreading following intrathecal
injection. Arrows indicate potential directions for protein flux from HAMC into the CSF or into
spinal cord tissue. (B) Tissue sampling paradigm for observing depth of dorsoventral
penetration. Spinal cord segments of 2 cm spanning the entire rostro-caudal spread of the
hydrogel were removed and then sectioned sagittally into 1mm thick sections. The most medial
two sections were then sliced longitudinally into 300 μm sections. These 1mm x 300µm x 2cm
sections were used to measure the dorsoventral penetration of FGF2 or PEG-FGF2 from the
dorsal aspect to the center of the spinal cord. ............................................................................... 57
Figure 3.2: FGF2 with free cysteine groups reacts with PEG-mal to produce single PEGylated
FGF2 and double PEGylated FGF2 (with some unreacted, non-PEGylated FGF2). ................... 58
Figure 3.3: Representative gel electrophoresis shows molecular weight ladder, control FGF2,
products of PEG-FGF2 reaction isolated from heparin binding column, and flow through PEG
from heparin binding column. ....................................................................................................... 59
Figure 3.4: Dose dependent activity of FGF2 (white bars) and PEG-FGF2 (black bars) show that
bioactivity based on cell density is maintained after PEGylation and subsequent purification.
(n=5, mean ± standard deviation are shown). ............................................................................... 60
xii
Figure 3.5: In vitro release of FGF2 (□) and PEG-FGF2 () from HAMC drug delivery system
shows diffusive release within 24 h in almost identical concentrations. (n=5, mean ± standard
deviation are shown). .................................................................................................................... 61
Figure 3.6: Cumulative comparison shows a greater amount of PEG-FGF2 () in the injured
spinal cord than FGF2 (□) when delivered from the intrathecal HAMC drug delivery system. B
At 3 h post-intrathecal injection, higher concentrations of PEG-FGF2 () than FGF2 (□) have
diffused into the cord. Lines indicate the predicted concentration profiles of PEG-FGF2 (solid)
and FGF2 (dotted). C At 6 h post- intrathecal injection, drug depletion from the hydrogel results
in a plateau of drug in the tissue, but elevated concentrations of PEG-FGF2 () are still observed
in the spinal cord. Asterisks indicate statistical significance at p<0.05. (n=5, mean ± standard
deviation are shown). .................................................................................................................... 63
Figure 4.1: (A) One-compartment and (B) two-compartment model of blood flow. ................... 72
Figure 4.2: (A) Representative arterial region of interest at T3 where a cross section of the entire
rat is shown. The box outlines the inset on right showing anatomical landmarks, including the
contrast filled catheter inside the jugular vein, the aorta (used for determining the arterial input
function), the trachea, and the intact T3 vertebra. The corresponding arterial input function is
shown below the inset. (B) A representative region of interest near the injury epicenter (T1/T2),
and the corresponding tissue uptake curve below. (C) A representative region of interest 0.75-1
mm rostral to epicenter (T1) where a laminectomy was performed and the corresponding tissue
uptake curve below. (D) A representative region of interest 1.8-2 mm rostral to the epicenter
(T1/C7) and the corresponding tissue uptake curve below. .......................................................... 75
Figure 4.3: (A) Spinal cord blood flow (SCBF) and (B) permeability-surface (PS) area products
nearest to injury epicenter at T1/T2 for: = FGF2/HAMC: = aCSF; = HAMC (n= 8, mean
± standard error of the mean are shown; letters indicate statistical differences as compared to pre-
injury values where a = aCSF, b = HAMC, c = FGF2/HAMC). .................................................. 79
Figure 4.4: (A) SCBF and (B) PS at 0.75-1 mm rostral to epicenter at T1 for: = FGF2/
HAMC; = aCSF; = HAMC (n=8, mean ± standard error of the mean are shown; letters
indicate statistical differences as compared to pre-injury values where a = aCSF, b = HAMC, c =
FGF2/HAMC). .............................................................................................................................. 81
xiii
Figure 4.5: (A) Spinal cord blood flow and (B) permeability-surface area products at 1.8-2 mm
rostral to epicenter at T1/C7: = FGF2/HAMC; = aCSF; = HAMC (n= 8, mean ±
standard error of the mean are shown; letters indicate statistical differences as compared to pre-
injury values where a = aCSF, b = HAMC, c = FGF2/HAMC; * indicates statistical significance
between aCSF and FGF2/HAMC groups at that time). ................................................................ 83
Figure 0.1: Protein release from HAMC is linearly related to time0.5 for >60% release. ........... 142
Figure 0.2: NBQX and solubilized nimodipine (with ethanol) release from HAMC. Inset shows
linear relationship with time0.5 for >60% of release. .................................................................. 142
Figure 0.3: Nimodipine release from HAMC for solid nimodipine particulates sonicated to
380nm and 900nm. Biphasic release profile was observed where fast release was observed in the
first 8 h (solid line = 380nm, dashed line = 900nm) and was likely due to soluble nimodipine
fraction saturating HAMC. Slow release was observed for 3 d (small dashed line = 380 nm,
large dashed line = 900 nm) likely due to fraction released as surface of particles were
solubilized into surrounding media. ............................................................................................ 143
Figure 0.4: Comparison of nimodipine solubility in MC and HA, with different nimodipine
particle sizes: (◆) 100-μm nimodipine particles in 0.25 wt% HA; and (□) 100-μm, (▲) 900-nm,
and (○) 380-nm nimodipine particles in 7 wt% MC. The upper shaded concentration range
indicates nimodipine solubility values achieved in 7 wt% MC, while the lower shaded range
indicates aqueous nimodipine solubility reported in the literature. ............................................ 144
Figure 0.5: (A) Longitudinal echogram imaging of the spinal cord where brightness indicates
density of the tissue. The bright spots show the dense bones of the vertebrae surrounding the
spinal cord; yellow line indicates where the laminectomy was performed to expose the spinal
cord. The spinal cord tissue can be observed between the vertebrae only below where the
laminae have been removed. (B) Arterial flow from vessel in spinal cord. (C) Venous flow from
vessel in spinal cord. ................................................................................................................... 145
xiv
List of Appendices
Appendix 1: Glossary 137 ..................................................................................................................
Appendix 2: Abbreviations 139 ..........................................................................................................
Appendix 3: Additional Data 142 .......................................................................................................
1
1 Background & Introduction
1.1 Rationale The central nervous system (CNS) is the integration center for signals that control bodily
functions, and these signals are transmitted through the peripheral nervous system. The CNS is
composed of the brain and the spinal cord, both of which are encased in bone as well as 3 semi-
impermeable meninges. Additionally, blood vessels in the CNS tightly regulate molecular
passage by means of the blood-brain and blood-spinal cord barrier. In traumatic spinal cord
injury (SCI), some or all of these layers of protection can be compromised, leading to loss in
limb, bowel, and sexual function, also making victims more susceptible to infections [1, 2].
Medical techniques are advanced enough to stabilize SCI patients immediately after injury,
giving these patients high life expectancies, yet no standard clinical treatment exists to prevent
the devastating loss of function. These patients permanently rely on wheelchairs or other devices
for locomotion, and require ongoing medical treatment. The resulting economic impact is
enormous on both the family of these patients and the government for medical care, with federal
estimates at $10 billion/year in the US and $1 billion/year in Canada [3]. The National Spinal
Cord Injury Statistical Center (NSCISC) estimates that there are more than 300,000 patients who
have suffered and survived an SCI in the US, with more than 12,000 new cases each year [4],
although a recent report suggests more than 1.2 million people live with SCI in the US [5]. The
economic and physical hardships that SCI patients endure emphasize the need for developing
new treatment strategies to improve the quality of life for these patients.
Current clinical treatment for SCI is intravenous injection of the steroid methylprednisolone
sodium succinate (MP). However, very high doses are required to elicit these responses [6, 7],
and is no longer standard practice since the limited functional benefit seen in clinical trials has
been debated [6]. Many drugs and proteins are under investigation for treatment of spinal cord
injury [8], and can generally be classified as either neuroprotective – those which limit neural
tissue loss after acute SCI – or neuroregenerative – those which promote neural tissue
regeneration after SCI. While intravenous injections are simple and the most widely used
method clinically, the high doses required to reach the injury site lead to unwanted systemic side
2
effects. Additionally, many molecules delivered in this manner either have a short systemic half-
life or cannot be transported across the blood-spinal cord barrier. To overcome these limitations,
our lab first developed an injectable collagen gel for delivery to the intrathecal space – the space
in which cerebrospinal fluid (CSF) cushions the spinal cord – thus circumventing the blood-
spinal cord barrier [9]. Subsequently, Gupta et. al. [10] characterized a fast gelling, injectable
biopolymer blend of hyaluronan and methylcellulose (HAMC) that improved some of the
qualities of the collagen gel, making it more suitable as an intrathecal drug delivery system. In
vivo results showed that this physical gel was biocompatible and safe for localized delivery to the
injured spinal cord. The research projects described in this document were undertaken to further
investigate the potential of HAMC as a delivery system for neuroprotective agents after
traumatic SCI.
1.2 Hypothesis and Objectives The hypothesis investigated in this body of work is that:
Localized delivery of neuroprotective agents with hyaluronan and methylcellulose will
mitigate tissue damage after spinal cord injury.
To test this hypothesis, two proteins were investigated as neuroprotective agents – erythropoietin
(EPO) and fibroblast growth factor 2 (FGF2). Three objectives were identified for this project as
described below, each one followed by a short summary of the supporting work presented in this
dissertation.
1. Characterize the in vitro and in vivo release of neuroprotective agents from HAMC
The release mechanism of EPO from HAMC was determined in vitro and is further described in
chapter 2. Both in vitro release and in vivo tissue delivery of FGF2 from HAMC was
investigated and improved by conjugation of poly(ethylene glycol) to FGF2 and is discussed in
chapter 3.
2. Determine the extent of in vivo neuroprotection achieved following localized delivery of
neuroprotective agents to the injured spinal cord
3
Neuron sparing achieved with EPO delivered locally from HAMC was compared to that of other
common delivery strategies and is further discussed in chapter 2. Spinal cord blood flow and
permeability in the spinal cord were assessed following localized delivery of FGF2 with HAMC
and is discussed in chapter 4.
3. Develop a live animal imaging strategy to assess spinal cord hemodynamics after
delivery of neuroprotective agents with HAMC
Methods for assessing functional spinal cord blood flow and permeability are limited, thus
computed tomography imaging was developed as a technique to assess spinal cord blood flow
and permeability following localized delivery of FGF2 with HAMC. This technique and the
models employed for it are presented in chapter 4.
1.3 Anatomy and Physiology of the Human Spinal Cord
The CNS is composed of the brain and spinal cord, both of which have multiple layers of
protection. The skull and the spinal column are the first layer of protection, providing a rigid
encasement for the fragile tissue within. Second, the meninges are membranes that are
selectively permeable to only very small molecules – cells, blood, and proteins are inhibited by
this fibrous membrane that envelops the CNS. The outermost layers of the meninges consist of
dense connective tissue and are known as the dura mater and the arachnoid membrane. The
innermost layer closest to the spinal cord is the pia mater, which is a highly vascularized, loose
connective tissue membrane. Between the pia and the arachnoid membrane is a space where
cerebrospinal fluid (CSF) flows and is known as the subarachnoid space or the intrathecal (IT)
space (Figure 1.1) [11]. Finally, the blood-brain and blood-spinal cord barrier (BSCB) restricts
the passage of substances into the CNS, allowing tight control of the extracellular environment in
the parenchyma.
The CSF cushions the brain and spinal cord from sudden movements, delivers nutrients and
removes waste for the CNS. It is an acellular solution that is mainly composed of salts, and is
produced in the choroid plexuses of the lateral, third and fourth ventricles of the brain. CSF
flows caudally and in a rhythmic pattern contributed by respiration, blood flow, and heartbeat at
4
a rate varying from -2cm/s to 8cm/s for humans [12]. The fluid circulates in the IT space around
the brain and the spinal cord before being reabsorbed via the arachnoid granulations in the
superior sagittal sinus and sinuses in the meninges of the spinal nerve roots [13].
Figure 1.1: Cross section of spinal cord showing the multiple layers of protection in CNS
(Copyright Michael Corrin).
There are several different cell types that make up the CNS, each of which tightly regulates and
maintains the extracellular environment. Neurons form the information conducting pathways
between the brain and the rest of the body. Glial support cells – oligodendrocytes, astrocytes,
and microglia – act to support neuronal function, regulate the BSCB, and protect the tissue from
infection. Neurons are polarized cells where dendrites receive electrical impulses from
preceding neurons, and axons transmit these electrical signals to successive neurons [11]. The
electrical signals are generated by changes in ion gradients across the cell membrane,
propogating down the axon to the synapse where the signal is passed to the subsequent neuron,
and this cycle is repeated until the signal reaches the target site or tissue. Neurons can be up to a
meter long, but must transmit these electric impulses very rapidly [11], thus can be insulated by
oligodendrocytes which form myelin sheath around axons. The spinal cord is comprised of the
white matter, densely populated with axons and dendrites, and gray matter, populated mainly by
cell bodies and the synapses which connect neurons [14]. The white and gray matter in the
5
spinal cord is segregated such that the butterfly shaped gray matter is surrounded by white matter
(Figure 1.1). Oligodendrocytes are found mainly in the white matter as anchorage and
maintenance cells, and astrocytes regulate the extracellular chemical environment in both the
white and gray matters [14]. Microglia are phagocytic cells that are normally dormant in the
CNS, but are activated to macrophages after injury, infection, or disease [14]. Ependymal cells
that line the central canal of the cord are considered an endogenous precursor cell population
with the potential to differentiate into several cell types in the CNS.
Three main cell types regulate the BSCB, acting in tandem to dynamically regulate ion balance,
facilitate nutrient transport, and prevent the entry of potentially harmful molecules. Endothelial
cells, pericytes and astrocytes comprise what is known as the neurovascular unit [15]. The
endothelium and abluminal pericytes that line blood vessels are surrounded by a thick basal
laminae composed mainly of collagen IV, and finally astrocytic endfeet ensheath these
capillaries. The endothelial cells are tightly connected by tight junction proteins which limit the
passage of molecules based on size, and it is thought that the astrocytic endfeet further limit what
passes into the parenchyma of CNS tissue [16].
1.4 Pathology and Prevalence of Spinal Cord Injury The National Spinal Cord Injury Statistical Center (NSCISC) estimates that there are more than
300,000 patients who have suffered from SCI in the US and Canada, with 12,000 new cases each
year [4]. A recent report by the Christopher and Dana Reeve Foundation estimates a far higher
incidence where over 1.2 million people in the US report being paralyzed due to a spinal cord
injury, and nearly 50% of those people reported “a lot of difficulty in movement” or “a complete
inability to move” [5]. The economic impact is estimated to be $10 billion/year in the US and $1
billion/year in Canada [3]. According to NSCISC estimates of SCI, 42 % are caused by motor
vehicle accidents, 27 % caused by falls, 15 % from acts of violence, 8 % from sports accidents,
and 8 % from other causes. More than 81% of SCI patients are male, and the average age at
injury is 40 years old [4]. Patients who survive the initial injury can have a normal lifespan, they
cannot however, lead a normal life. They and their families are burdened with lifetime
healthcare and living expenses of up to $2 million, yet SCI patients often have difficultly
integrating into the workforce to allay these costs.
6
Table 1.1: Secondary injury timeline
Bunge et. al. [17] has provided a simple classification
method for SCI based on gross findings: solid cord injury,
contustion/cavitation, laceration, and massive compression.
Complete cord transection is a paradigm used in
experimental animal models, but is not commonly found in
humans [18]. Compression of the cord can be caused by
dislocation or fracture of the vertebra in the spine, and is the
most common form of SCI [3]. The anatomic level and
severity of injury determine the extent of the resulting loss
of function. An injury at a given vertebral level impacts all
levels below the injury. The cervical levels and the upper
thoracic region control the upper extremities and respiration.
A patient that survives an SCI in this region may require a respirator and could be a quadriplegic.
If an injury is sustained in the lower thoracic region or the upper lumbar region, they may be a
paraplegic without bladder or bowel function. Injuries in the lower lumbar region would also
affect the posterior of the lower extremities and may necessitate the use of crutches or other
devices to assist them in locomotion. Since no clinical treatment is effective at limiting the
tissue degeneration or regenerating functional CNS tissue, these patients live with these
limitations for the rest of their lives.
The primary insult to the spinal cord in a compression injury is the mechanical trauma that
crushes the tissue. While this causes immediate cell death and vascular leakage, these impacts
cause subsequent pathological events that continue to occur well after the injury, and this
cascade of events is termed the secondary injury. The secondary injury has a significant impact
on the outcome of the tissue, thus has serious implications on the potential for functional
recovery. The timeline of events discussed here is indicative of the secondary injury process in
humans (see Table 1.1). The early phase, beginning within hours of the injury and lasting for ~2
days post-injury, is characterized by edema, hemorrhage, inflammation, vascular changes, as
well as neuronal and myelin changes [18]. Edema is fluid accumulation in the extracellular
space that leads to swelling, and can be caused by hemorrhage and leakage of plasma fluid into
the parenchyma as a result of vascular rupture. The inflammatory response is a complex process
Acute (hours to ~2 days) Edema Ischemia Lipid peroxidation Free Radical formation Hemorrhage Inflammation Neuronal/Axonal changes Demyelination Intermediate (days to weeks) Microglial responses Astrocytic responses Vascular responses Late (weeks to months/years) Wallerian degeneration Scar formation Cavity formation/elongation
7
involving cellular and chemical mediators. Neutrophils and other immune regulators begin
infiltration within hours of injury, and may persist for 3-4 days. Relative to the neutrophilic
response in other tissues this is quite short, but they release free radicals that contribute to tissue
damage [19]. Microglia native to the CNS start to become activated macrophages early post-
injury, and begin phagocytosis of cellular debris. The necrotic and apoptotic neurons from the
primary and secondary injury flood the extracellular space with intracellular ions,
neurotransmitters, and other cell regulators, leading to a feedback loop in which further neuronal
cell death is propagated. Oligodendrocytes forming the myelin sheath swell after injury and
eventually are fragmented, in turn contributing to the hostile extracellular environment. These
excess neurotransmitters, reactive oxygen species, ammonia, and free radicals, among others
[18] accumulate in the extracellular space, disrupting homeostasis of a much larger region than
that affected by the primary injury.
The intermediate phase of the secondary injury occurs within days to weeks after the initial
injury [18]. The activated macrophages continue to phagocytose cellular debris, secreting factors
that are considered harmful, such as nitric oxide, glutamate and proteases, although they also
release protective cytokines, such as interlukin-1β [20] and tumor necrosis factor-α [21]. Several
days after the initial insult, astrocytes will undergo hypertrophy and begin to swell. These are
considered activated astrocytes, and after 2-3 weeks will begin to expand and send out thick
processes, ultimately forming what is termed the glial scar, or the astrocytic scar [18]. This scar
was once believed to be a physical barrier for regeneration, but more recent reports have shown
that by the time the scar fully forms, regenerating axons should be beyond the scar [18].
Additionally, astrocytes are involved in restoration of the extracellular ionic environment, uptake
of neurotransmitters, free radical scavenging, and provision of nutrients, among others [22].
They are also sources of growth factors and neurotrophic compounds which can provide
neuroprotection and promote recovery. Within a week after SCI, spontaneous angiogenesis is
observed in the spinal cord [18] and restoration of the BSCB begins to occur, which are thought
to be regulated by factors secreted from astrocytes [23, 24]. In this intermediate phase, it is
evident that while macrophages, astrocytes, oligodendrocytes, and neurons each play an
important role in restoring homeostasis within the spinal cord, each of these can also have
negative impacts on the tissue. Thus, designing effective strategies for neuroprotection and
neuroregeneration that can lead to functional recovery is much more complex than simple
8
inhibition or down-regulation of a specific cellular response, suggesting the need for a
combination drug or protein therapy.
In the late stages of secondary injury, further neuronal death may occur through a process known
as Wallerian degeneration. This is characterized by anterograde degeneration of axons and the
myelin sheaths, a process which can occur for over a year [18]. The astrocytic scar also is
stabilized and the astrocytic processes become tightly interwoven with extracellular matrix
proteins such as chondroitin sulfate proteoglycans [25, 26] thereby stabilizing the early effects of
the secondary injury such as hemorrhage, edema and cellular debris and leading to a fluid-filled
cavity or cyst in the spinal cord. While most factors are cleared from the fluid over time, this
fusiform cavity provides no scaffold or substrate for future neural regeneration. This cavity
persists for the lifetime of the patient, and in some occasions a pressure filled cyst called a syrinx
elongates, causing further damage to the spinal cord [18]. Syrinx formation is often associated
with arachnoid or fibrous adhesions that prevent clearance of fluids, and can lead to conditions
such as syringomyelia [27].
The early cell death and accumulation of toxic factors which progressively lead to further
degeneration emphasizes the importance of neuroprotective strategies. Reducing tissue loss from
necrosis and apoptosis may lead to better tissue and functional outcomes [28]. Neuroprotection
in the early stages of secondary injury may also lead to an environment that is more conducive to
regeneration.
1.5 Animal Models of Spinal Cord Injury The timeline and the events discussed above are with respect to the pathological changes seen in
the human spinal cord. Because the overall goals for drug delivery to the spinal cord are for
human use, it is important to understand these conditions. However, for the in vivo testing phase
of materials and drugs that will be used in drug delivery systems, animal models are very
important and are commonly used. There are several animal models that can be used for testing
biocompatibility and efficacy of the DDS, as well as various injury models. Rodent models are
the most commonly used for initial animal studies, but studies have also been done in guinea
pigs [29, 30], canine [31], feline [32], murine [33], and primate [34, 35] subjects as well. Rats
9
are the most common small animal model for traumatic SCI because the morphological,
biochemical, and functional changes that occur after SCI are similar to those seen in humans,
although differences are likely at the molecular level [36-38]. Female rats are preferred because
manual bladder expression after SCI is required since this function is affected by SCI at almost
all vertebral levels normally tested in vivo.
Four main injury models have been developed to study the mechanisms of SCI: hemisection
[39], transection [40], compression [34, 41-43], and contusion [44, 45]. A more complete review
of these models can be found in [46], and are illustrated in Figure 1.2. Hemisection models are
done by either cutting a dorsoventral or mediolaterial section from the cord. These are useful
models since the injury site is reproducible and the contralateral or uninjured segment of the cord
can be used as a control in some cases. However, the extent of tissue that is removed is difficult
to control, sometimes leading to high variability in tissue and functional recovery. The complete
transection model, which involves opening the dura and transecting the entire spinal cord, has
advantages of certainty about the extent of injury, but the severity of the injury makes animal
care more difficult [47]. These two previous models are particularly useful for regenerative
strategies because there is little question between spared and regenerating axons. However,
compression and contusion models are more relevant to human injuries, and are ideal for
studying mechanisms of neuroprotection [35, 42]. Contusion injuries in humans occur from
blunt trauma where an external force causes a quick but considerable blow to the cord. Some
well established models in rats and mice for contusions are the New York University Impactor
[44] and the Ohio State Accelerator [45], where a weight is dropped onto the exposed spinal cord
to apply a quick and measurable force. These models have extensive variability between animals
unless properly controlled by computer automation [48]. These types of injuries tend to damage
the large myelinated axons and spare the smaller axons [49]. Compression injuries, the most
common form in humans, are caused by prolonged pressure on the spinal cord tissue producing
blockage of blood vessels and axonal conduction [49]. Two models of compression used to
induce SCI are subdural balloon inflation [50] and the modified aneurysm clip [34, 41-43]. In
the clip compression method, the cord is compressed for 1 min by a clip calibrated to a specific
closing force, and is reasonably reproducible. With this model, mechanisms of secondary injury
and the effects of neuroprotective drugs can be studied. Since compression injuries are more
clinically relevant and the clip compression model has been well characterized [43, 51], all
10
animal studies discussed in this dissertation were done with female rats using the clip
compression method for inducing traumatic SCI.
Figure 1.2: Illustration of leading models of spinal cord injury (Copyright Michael Corrin).
1.5.1 Measurement Techniques for Spinal Cord Hemodynamics
Grossly injured vessels from the primary injury cause immediate hemorrhage and progressive
edema in the parenchyma of the spinal cord; both spinal cord blood flow (SCBF) and the BSCB
are compromised by these events. The accumulation of plasma constituents in the cord alters the
tightly controlled chemical environment of the CNS, and ischemia and hypoxia are ongoing until
blood flow is resumed. Breakdown of the BSCB is manifested by permeability to serum proteins
within 3 h in areas adjacent to the injury epicenter [52], and SCBF is reduced within 2 h [53].
While hypervascularity can occur after injury to restore blood flow [54], permeability of the
BSCB remains quite high even at 28 d post-injury [55]. This quantitative spatial analysis of the
BSCB was performed by injecting [14C]-α-aminoisobutyric acid intravenously at a time of
interest post-injury followed by rapid harvesting of the spinal cord. Segments were then
sectioned and optical density of autoradiograms measured densitometrically were used to assess
the amount of tracer present in the spinal cord tissue. Another study showed that the BSCB is re-
established within 14 d post-injury [52], assessed by injecting horseradish peroxidase into the
11
bloodstream and analyzing tissue sections with electron microscopy. While these studies were
useful to demonstrate the dynamic changes of the BSCB, they illustrate that permeability
measurements are highly dependent on the extent of the injury and the size of the molecule used
to assess leakage. Recently, magnetic resonance imaging has been utilized to assess
permeability, but thus far has been limited to qualitative assessments [56, 57].
SCBF has been measured using several techniques, but most are invasive or limited to select
regions or time points. Kety and Schmidt first proposed gas clearance techniques for measuring
SCBF, and developed models based on Fick’s principles of diffusion [58, 59]. In these
techniques, an inert gas is breathed in through the nose or mouth, and allowed to equilibrate with
the tissue of interest. Because most gases are quite permeable to the BSCB, the gas will
equilibrate within minutes. The gas flow through the nose or mouth is then stopped, and a probe
placed directly into spinal cord tissue measures gas pressures as the gas clears from the tissue or
it can also be measured by continuous sampling of uncontaminated venous blood coming from
the tissue. Gases that have been used for this purpose are mainly xenon [60, 61] and hydrogen
[53, 62]. While this technique allows for repeated measurements while the probe is in place, the
probe insertion causes trauma and data is only indicative of local measurements rather than a
composite of the whole cross sectional segment of the cord. Radioactive gases or particles have
also been used in combination with autoradiography to measure SCBF quantitatively. In these
methods, the gas or particulate tracer is abruptly introduced into the arterial blood immediately
prior to harvesting the tissue. After sectioning, autoradiographs of frozen tissue are used to
measure concentrations of the gas or particulate densitometrically. Trifluoroiodomethane
labeled with 131I was the first to be used with this method [63], but [14C]antipyrine developed
later is more commonplace for these methods [50, 64]. These measurements allow only a single
measurement from an animal since tissue harvesting is a necessity, and variability between
animals can be problematic. More recently, the Laser Doppler technique has been used, but
since it requires the probe to be on the surface of the spinal cord, it is limited to a relatively short
timeframe after the animal has undergone the injury or a terminal time point [65]. Permeability
has only been assessed quantitatively using histological and immunohistochemical techniques on
sectioned tissue samples. While recent studies have attempted to utilize magnetic resonance
imaging (MRI) techniques [56, 57], measurements were only qualitative.
12
Recent advances in MRI and computed tomography (CT) imaging techniques have produced
better resolution of tissues than in previous decades. In addition, dynamic contrast enhancement
(DCE) with contrast agents injected into the bloodstream has allowed measurements of blood
flow and permeability in soft tissues with these techniques [66-74]. CT has been used to
investigate both hemodynamic changes in growing tumors [67, 69, 71-74] and cerebral blood
flow [67, 73]. These techniques are preferable since they are minimally- invasive techniques and
allow repeated measurements to be taken over a relatively large tissue range in the same subject.
In these methods, a relatively impermeable tracer, such as iodinated compounds or gadolinium
particles are injected into a vein close to the heart while fast acquisition CT or MRI scans are
being performed. Two main models are used to mathematically determine both blood flow and
permeability parameters from the scans: the modified Toft’s model [75, 76] and the adiabatic
tissue homogeneity model [77, 78]. Generally, these are quite similar models, but the adiabatic
tissue homogeneity model requires higher signal to noise ratios to generate accurate data. One
study of SCBF in humans has been reported using dynamic perfusion CT [66]; however, these
have not been replicated in small animal studies. Techniques such as CT and MRI which could
be performed in both humans and small animals could prove very beneficial in comparing
outcomes between preclinical animal studies and clinical trials of treatment strategies.
1.5.2 Motor Function Outcome Measures
Since animal models are common for testing SCI treatments, several tests have been developed
to assess functional recovery following SCI. Because complete high cervical injuries require
extensive animal care and often have higher mortality rates, forelimb testing is mainly used for
some hemisection models. For complete SCI models which are mainly done in thoracic and
lumbar regions, such as the compression model, hindlimb testing is most common. The inclined
plane test [79], narrow beam walk [80], grid walk test [81, 82], footprint analysis [83], and the
Basso, Beattie, and Bresnahan (BBB) open field locomotor test [84] are primarily used, several
of which were tested in both a hemisection and contusion injury model in Metz et. al. [85]. In
any SCI model, a variety of tests are often performed, however, the BBB test is most prolific and
is often used as a baseline of comparison between studies performed in different laboratories. It
is a 0-21 point scale based on anatomical markers and specific locomotion behaviors which are
13
observed for rats allowed to freely walk in an open field environment. These are summarized in
Table 1.2 and further described below.
Table 1.2: Basso, Beattie and Bresnahan (BBB) locomotor score rankings.
Score Characteristics 0 No observable hindlimb (HL) movemements 1 Slight (≤ 50 %) movement of one or two HL joints
2 Extensive (>50 %)movement of one HL joint and possible slight movement of one other joint
3 Extensive movement of two HL joints 4 Slight movement of all three HL joints 5 Slight movement of two HL joints and extensive movement of third HL joint 6 Extensive movement of two HL joints and slight movement of third HL joint 7 Extensive movement of all three HL joints
8 Sweeping (rhythmic extension of three HL joints) with no weight support OR Plantar placement with no weight support
9 Plantar placement with weight support in stance only OR Occasional, frequent or consistent weight supported dorsal stepping and no plantar stepping
10 Occasional (>5 % and ≤50 %) weight supported steps with no forelimb-hindlimb (FL-HL) coordination (one HL step per FL step, alternating HL steps)
11 Frequent (51-94 %) to consistent (95-100 %) weight supported steps and no FL-HL coordination
12 Frequent to consistent weight supported steps and occasional FL-HL coordination 13 Frequent to consistent weight supported steps and frequent FL-HL coordination
14 Consistent coordinated plantar stepping, predominant paw position is rotated at initial contact and liftoff; Frequent plantar stepping, consistent FL-HL coordination, and occasional dorsal stepping
15 Consistent coordinated plantar stepping, no or occasional toe clearance during forward libm advancement, predominant paw position as parallel to body at initial contact
16 Consistent coordinated plantar stepping, frequent toe clearance, and predominant paw position is parallel to body at initial contact and rotated at liftoff
17 Consistent coordinated plantar stepping, frequent toe clearance, predominant paw position is parallel to body at initial contact and liftoff
18 Consistent coordinated plantar stepping, consistent toe clearance, predominant paw position is parallel to body at initial contact and liftoff
19 Consistent coordinated plantar stepping, consistent toe clearance, predominant paw position is parallel at initial contact and liftoff, tail is down part or all of the time
20 Consistent coordinated plantar stepping, consistent toe clearance, predominant paw position is parallel to body at initial contact and liftoff, tail is consistently up, and trunk instability
21 Consistent coordinated gait, consistent toe clearance, predominant paw position is parallel at initial contact and liftoff, tail consistently up, and consistent trunk stability
14
Various regions of the scale reflect the severity of injury and the degree of recovery [86].
Generally, the scores from 0-8 focus on simple limb movements that are displayed in the initial
stages of recovery, and animals that have an extensive injury, such as transection or massive
compression of the spinal cord, often never recover beyond this range. The 9-14 range describes
weight support and coordination behavior seen in the intermediate stage of recovery. Animals
with moderate compression or contusion injuries often plateau in this range 2-4 weeks post-
injury. The 15-20 range encompasses more subtle movements like coordination, foot placement,
and tail and trunk posture in walking behavior. This range is useful for assessing locomotor
changes after very mild injuries or for assessing the safety of a treatment in uninjured animals.
Methods such as the grid walk test are used to examine differences in more fine motor skills,
such as coordination in stepping and foot placement. In this type of test, animals are trained to
walk across a grid and footsteps and footfalls through the grid indicate improvement in fine
motor skills where fewer footfalls suggests better motor control.
1.6 Treatment Strategies for Spinal Cord Injury Repair Currently, there are no effective strategies for repairing the spinal cord after traumatic injury, but
many neuroprotective and neuroregenerative strategies are under development. Some of these
involve drug delivery systems [10, 87], cellular delivery systems [88-90], modification of drugs
to cross the blood-brain barrier for intravenous delivery [91, 92], scaffolds for promoting axonal
outgrowth [93, 94], entubulation of the spinal cord to induce tissue bridge formation [90, 95],
and transplantation of adult or fetal neural tissue [96]. The discussion here will focus on
clinically tested treatments for SCI and potential neuroprotectants that can be delivered locally
with HAMC.
1.6.1 Clinical Trials and Treatments
Immediately following SCI, the priority is stabilizing the vital signs of a patient. The location,
degree, and type of injury dictate the type of treatment a patient receives. Strong experimental
evidence as well as class II and III clinical data suggest that surgery should be done to remove
bone fragments and decompress the spinal cord for acute, non-penetrating SCI [97, 98].
However, this is performed at the discretion of the physician, and recent recommendations
15
suggest this should be done 8-24 h post-injury [99]. The only drug treatment used is intravenous
administration of MP, as previously discussed. The mechanistic rationale for use of this
corticosteroid is to reduce edema and ensuing swelling that occurs after primary injury. MP also
has a strong antagonistic effect on lipid peroxidation, which occurs extensively after injury
[100]. In parallel, MP has been shown to inhibit ischemia, support aerobic energy metabolism,
reduce intracellular overload of calcium, and attenuate calpain-mediated neurofilament loss [28].
However, clinical trials of MP showed that high doses were required to benefit patients, yet the
functional impact was minimal [101-104]. Concerns surround the small sample size as well as
the non-standardized surgical protocols [6, 7], and currently is not used in many Canadian and
European countries, although it is still widely used in the United States.
Several other drug treatments using systemic delivery have undergone clinical trials, such as
tirilizad mesylate [105], thyrotropin releasing hormone [106], monosialotetrahexosylganglioside-
1 (GM-1) [107, 108], naloxone [103, 104], gacyclidine [109], and nimodipine [110]. GM-1
showed faster achievement of peak functional recovery, but did not show an increase in
functional improvement. In animal studies, Naloxone demonstrated improved spinal conduction
and a reduction of edema, but in clinical trials failed to show therapeutic effects [103, 111].
Tirilizad mesylate possesses antioxidant properties and stabilizes cell membranes by limiting
lipid peroxidation [112, 113]. However, it did not show functional benefit over placebo controls
in Phase III trials, and is not used as a clinical treatment [105]. Molecules that were tested in
small clinical trials were gacyclidine [109], thyrotropin releasing hormone [106], and nimodipine
[110], yet none showed significant beneficial results, thus were not further pursued. Recent trials
are underway for Cethrin, riluzole, and anti-Nogo A, however, these results are not yet available
[114]. The US Food and Drug Administration did not approve the delivery of Anti-Nogo A with
minipumps since they can cause scarring and compression of the cord [115], so these studies are
currently being done in Europe and Canada only [114]. No clinical trial has shown impairment
due to any drug treatment, however, the lack of effective treatment resulting from so many
clinical trials underscores the need for more research to develop effective treatment strategies
with the goals of improving functional outcome for SCI patients.
16
1.6.2 Experimental Neuroprotective Agents
Neuroprotective agents are those molecules which have an impact on neuronal survival, directly
or indirectly, following an injury or biochemical insult. After traumatic SCI, most cells damaged
from the primary injury cannot be spared; however, manipulation of the biochemical events
which occur in the secondary injury may allow for greater survival and tissue sparing in the
penumbra surrounding the injury. A plethora of literature exists for neuroprotective agents that
have been tested in vivo for treatment of SCI, as reviewed in [8, 116]. Table 1.3 summarizes
some of these neuroprotective agents, identifying the benefits and limitations of each. This table
however, should not be considered an exhaustive list, but only a guide to the molecules that have
been published to date. Each molecule listed is classified based on the secondary injury pathway
affected, although some may have a variety of actions. Delivery with HAMC was expected to
result in the greatest benefit by delivering proteins that possessed several molecular targets and a
wide range of beneficial actions in vivo. For this reason both erythropoietin (EPO) and
fibroblast growth factor 2 (FGF2) were selected as good candidates to deliver with HAMC; the
actions of each are further described in subsequent sections. Localized delivery of EPO with
HAMC was investigated in comparison to conventional drug delivery strategies since it is a
neuroprotective molecule that crosses the blood-spinal cord barrier. Since FGF2 does not cross
the blood-spinal cord barrier, it was investigated with respect to intrathecal controls only. Both
proteins provided a unique perspective on the performance of HAMC as an intrathecal drug
delivery system.
17
Table 1.3: Drugs for SCI treatment in cinical or experimental studies indicating potential
benefits and limitations Class of drug
Drug Potential/Perceived Benefit Potential Limitations
Steroids Methylprednisolone Sodium Succinate (MP) [101, 103, 105]
Phase III clinical trials indicate minimal functional improvement; Prevents lipid peroxidation
Functional benefit is limited; High doses increase wound infections, pneumonia, and respiratory problems
Tirilizad mesylate (TM) [105]
Does not activate glucocorticoid receptors, so may not have negative side effects of MP
Showed no improvement in functional recovery during clinical trials
Gangliosides GM-1 [107, 108] Prevents lipid peroxidation and may enhance neuronal survival
Only improved recovery for lower extremities, but effect was not significant for higher SCI
Opiate blockers
Naloxone [103, 104, 117]
Improves spinal cord conduction and reduces edema
Failed in clinical trials to show therapeutic benefit and functional recovery
Glutamate receptor antagonists
NBQX [118] Attenuates neurotoxicity, and has been shown to be safe in vivo
Large doses required to provide neuroprotection can globally affect nervous system
Gacyclidine [109, 119, 120]
Improved functional, histological and electrophysiological status after contusion
Long term functional benefit was lacking in clinical trials; Global affect on nervous system
Taurine [121] Modulates intracellular calcium homeostasis; prevents neurotoxicity in vitro
Globally affects glutamate receptors in the nervous system
mGluR antagonists [122]
Decreases excitotoxicity; Improved locomotor function and spares gray matter
Require spinal delivery for beneficial affect
Ca2+ Nimodipine [110] channel blockers
Prevents excitotoxicity and vasospasm induced ischemia
Clinical trials failed to show neurological benefit
Sodium channel blockers
Tetrodotoxin [123] Reduces excitotoxicity from ion influx; Axonal loss attenuated after contusion in rodents
Global affects on nervous system for intravenous administration
Riluzole [124, 125] Reduces excitotoxicity from ion influx; White and gray matter sparing seen after contusion in rodents; approved by FDA for ALS therapy
Global affects on nervous system for intravenous administration
Calpain antagonists
CEP-4143 [126, 127]
Prevents destabilization of cytoskeleton and myelin protein catabolism; In vivo tests show decreased neurofilament breakdown and improved locomotion
Requires invasive subdural delivery to show benefit
Immune blockers
Cyclosporin-A [128] Suppresses the immune system, prevents lipid peroxidation better than MP; reduces demyelination and neuronal cell death and improves motor outcomes
Globally inhibits immune system and could cause increased infections and complications
Anti-Integrin mAbs [129, 130]
Prevents leukocyte and neutrophil tethering and infiltration; Enhanced neurological function after intravenous administration
Globally inhibits immune system and could cause increased infections and complications
Minocycline [131, 132]
Antibiotic could decrease inflammation; Decreases oligodendrocyte apoptosis and microglia activation, reduce lesion size and improve neurologic deficit
Reduced microglia activation may not be beneficial because macrophages phagocytose debris and secrete growth promoting factors
Polymers Polyethylene glycol (PEG) [133, 134]
Dehydrates membranes and prevents lipid peroxidation; Has shown functional benefit with severe SCI models
Inappropriate neuronal connections may form since PEG is a fusogen
18
Proteins and peptides
Anti-NOGO-A [135-137]
Targets myelin growth inhibitors; Prevents demyelination and promotes axonal survival
Very specific to oligodendrocytes, so would be best in combination with other neuroprotective agents
Basic Fibroblast Growth Factor (bFGF/FGF-2) [87, 138, 139]
Angiogenic factor; Has been co-delivered with EGF to stimulate proliferation of ependymal cells
Has not been extensively studied in spinal drug delivery; may cause tumor formation
Brain Derived Neurotrophic Factor (BDNF) [140]
Promotes survival of sensory and motor neuron populations; Promotes axonal regeneration
Controversial because other labs report no neuro-protective or regenerative affects
Epidermal Growth Factor (EGF) [87, 141]
Stimulates ependymal cell proliferation; Co-delivery with bFGF may stimulate differentiation
Used mainly for stimulating differentiation of neural stem cells
Erythropoietin (EPO) [49, 142]
Reduces lipid peroxidation, inflammation and apoptosis; Increased functional benefit and reduced lipid peroxidation in vivo
May increase hematocrit when administered systemically which can amplify SCI
Neurotrophin-3 (NT-3) [143, 144]
Mediate apoptosis in development; Promotes regeneration of axons after SCI
Optimal affects seen with long term delivery; cannot cross BSCB
Thyrotropin releasing hormone (TRH) [106, 145]
Antagonizing effect on endogenous opioids, platelet activating factor, and exitatory amino acids; Promising results seen for clinical trials
Clinical trials showed improved recovery, but had small patient population and lacked long term data
z-VAD-fmk Tripeptide caspase inhibitor [146, 147]
Caspase-3 inhibition prevents apoptosis when administered topically on dura following SCI; May show increased benefit with co-delivery of calpain inhibitors
Some studies showed benefit only with pre-treatment of drug, others showed no benefit after SCI
Rho pathway antagonists
Y-27632 [148, 149] Rho inhibition promotes axon sprouting; demonstrated axonal sprouting on inhibitory substrates when present
Regenerative effects are greater for C3 transferase
Cethrin (C3 transferase with transport sequence) [150]
Promotes axon sprouting, improves blood flow to ischemic regions; Phase I clinical trial for extradural administration was completed
Delivered epidurally from fibrin, limiting amount diffusing into spinal cord
1.6.2.1 Erythropoietin
Erythropoietin is a highly glycosylated 30kDa hormone that acts as the primary regulator of red
blood cell production in mammals [151]. The peptide moiety is 165 amino acids with 4
glycosylation sites; the average carbohydrate content is ~40%, which plays an important role in
its stability, solubility and biological activity [152]. In addition to its hematopoietic properties, it
also plays a role in central nervous system development [153-155]. Recent reports show that
EPO and its receptors are upregulated in the brain and spinal cord within minutes to hours after
an ischemic injury [156, 157]. The exact neuroprotective mechanisms of EPO are unclear; most
studies focus on its role in limiting ischemia in the CNS, but some studies show that it has
neurotrophic properties as well [158, 159]. In cultured rat cortical neurons, EPO has been shown
19
to reduce hypoxic [160] and glutamate [155] induced cell death. Tissue loss is decreased when
EPO is administered following ischemic damage [161], and it is thought to reduce inflammation
[162, 163] and spare white matter tracts in the spinal cord [49, 164]. EPO is thought to have a
proliferative effect on both oligodendrocytes and astrocytes, which potentially have an indirect
effect on sparing neurons after ischemic injury [156, 159, 165, 166]. The proliferative and
antiapoptotive effects of EPO are thought to act through the JAK2/STAT3 and the PI3K/AKT
activation pathways[158]. Positive functional recovery has been observed in stroke models
following intravenous EPO administration, which is attributed to the neuroprotective effects of
the glycoprotein which has been shown in hypoxic and ischemic neonatal and adult rat brains
[167-170]. In ischemic spinal cord injury, EPO delivered systemically reduced apoptosis of
motor neurons[161]. In a compressive SCI model, increased motor score and reduced ischemic
injury was observed [171], supported by another study with contusive SCI [49], where
significant functional recovery was observed. These studies showed evidence that localized
delivery of EPO would limit tissue damage and achieve significant neuroprotection after SCI.
Since EPO can cross the BSCB [172], it was an ideal neuroprotective molecule to test with the
HAMC delivery system to determine if localized delivery would be more beneficial than
systemic. Additionally, previous studies in the spinal cord showed that multiple systemic
injections of EPO was not better than a single immediate injection, suggesting that the
neuroprotective actions occurred very early in the secondary injury phase [49], which was an
optimal time frame for delivery from HAMC.
1.6.2.2 Fibroblast Growth Factor 2
Fibroblast growth factor 2 (FGF2) is a protein that exists in several forms of 18, 22, and 24 kDa,
all of which are highly regulated in vivo [173]. The 18 kDa FGF2 is the most ubiquitous, and
sequence homology is very high (>90%) across a wide range of species [174]. It contains a large
number of basic residues (pI=9.6) and 4 cysteine residues, none of which participate in disulfide
bonding, unlike most other proteins [175, 176]. It has 2 binding sites for FGF receptors (which
bind to other FGF family proteins as well) and also has a high binding affinity for heparin and
heparin sulfate proteoglycans [177]. Heparin acts to stabilize the tertiary structure of FGF2 and
enhances its bioactivity [178, 179]. FGF2 is best known for its trophic properties in promoting
20
angiogenesis [180-182], as well as the capacity for promoting differentiation of stem cells [183].
It is expressed in the CNS, and following brain injury [184] and SCI [185, 186], it is upregulated
although still tightly regulated [173]. FGF2 has also been shown to be neuroprotective [187,
188] and produce functional recovery in experimental models of SCI [138, 139]. It stimulates
endogenous ependymal cells to proliferate when co-delivered with EGF into either the ventricles
[189]or into the intrathecal space [87]. Ex vivo experiments show that it can modulate the
number of blood vessels in the CNS, and induce experession of tight junction proteins,
suggesting that it might reduce permeability of the blood-brain barrier in CNS lesions [190, 191].
FGF2 has also been used to limit ischemic CNS injuries [187, 192], and has been tested
clinically for the treatment of stroke [193]. Because this molecule is so ubitquitous and has
diverse effects, a single mechanism of action is unlikely. It is thought to promote angiogenesis
primarily by inducing the proliferation of endothelial cells , and also by upregulating the
expression of vascular endothelial growth factor (VEGF) [194]. FGF2 acts synergistically with
platelet derived growth factor to induce vessel maturation [195]. It is thought to reduce
permeability of the blood-brain barrier by upregulating tight junction proteins that limit passage
of molecules across the endothelium in the brain [190, 191].
These studies suggest that FGF2 will provide neuroprotection, stimulate endogenous stem cells
in the ependyma, and promote angiogenesis, perhaps limiting the detrimental effects of
secondary injury. Notwithstanding these obvious benefits, FGF2 does not cross the intact
blood-brain barrier [196], and is known to be mitogenic and upregulated in cancer [182],
reflecting the limitations of systemic administration [197]. Thus FGF2 was an excellent choice
for local delivery using HAMC after SCI.
1.6.3 Routes of Drug Administration
Traditional routes of drug delivery include intravenous bolus injections and oral delivery. Oral
delivery is disadvantageous for SCI since it must first be absorbed through the gastrointestinal
tract into the bloodstream, thus extending the time before treatment can reach the injury site.
Intravenous delivery is a simple method (Figure 1.3), but has several major limitations. Large
doses are required so that an adequate dosage will reach the target tissue, potentially causing
negative systemic side effects or toxicity to other tissues. Many drugs have short half-lives in
21
blood, and thus must be infused continuously. Additionally, the BSCB is a selective barrier that
normally keeps drugs and proteins out of the CNS, and although it is compromised at the site of
injury, several studies have shown that blood flow is severely decreased SCI, as previously
discussed. Systemic delivery is therefore not optimal for SCI drug treatments.
Localized spinal drug delivery has obvious benefits, but does not lack limitations. Epidural
delivery, into the soft tissue between the vertebrae and the dura surrounding the spinal cord, can
be administered much closer to the site of injury. Toxicity and side effects are reduced, but
transportation across the meninges and through the pulsating CSF is still required to reach the
tissue. Intrathecal and intramedullary methods are much closer to the site of injury and do not
require passage across the BSCB, but are more invasive (Figure 1.3). Intramedullary methods
involve injections directly into the cord, which could cause more damage to the injured cord
[198]. Osmotic mini-pumps have been used for intrathecal delivery into the CSF, and have
shown some promising results for delivery of analgesics (Figure 1.3) [199]. Unfortunately, the
implanted catheters can compress the spinal cord and are prone to infection since the dura may
not seal around the catheter insertion site. Additionally, the catheters can cause scar formation in
the intrathecal space, leading to catheter blockage, thus impeding drug flow [115]. Both
intrathecal and intramedullary delivery are invasive since the dura must be breached, potentially
resulting in long term leakage of CSF which causes massive headaches and pain. However,
intrathecal delivery may be an optimal route for delivery of some neuroprotective agents since
they can be delivered as close to the injury as possible without impinging on the injury site itself
and have the least systemic side effects. For acute SCI patients undergoing decompressive
surgery, the dura is exposed at the site of injury, providing a unique opportunity for treatment in
this manner. Intrathecal delivery provides direct contact with the spinal cord tissue, and only
requires the drugs to traverse the pia mater, which is more permeable than both the dura and
arachnoid maters [200].
22
Figure 1.3: Illustrations of drug delivery routes: (A) Intravenous [201], (B) intrathecal
bolus via lumbar puncture [202], and (C) implanted minipump (copyright Michael Corrin)
1.7 Polymers for Drug Delivery
Polymers, composed of identical repeating units, have emerged as useful materials for drug
delivery because their properties can be tailored for biological applications. They can act as
reservoirs for biomolecules and release can be controlled with specific material modifications.
Polymers can also be used to stabilize proteins and prolong the half-life of molecules or enhance
their distribution through tissue. Because polymers can have very different properties, such as
stiffness, porosity, and water content, drug delivery systems can vary widely from solid reservoir
systems [115], to sol-gel systems [203], films or wafer systems [204-206], or injectable flow
systems [9, 10, 87]. Although polymers are not without limitations, such as material-tissue
interactions or immunogenic responses, they are optimal biomaterials for drug delivery.
Drug release from polymer systems can be achieved through many approaches. Bulk drug
release that is diffusion mediated is the simplest method. Degradation mediated release can be
achieved by either entrapping a drug in a polymer with a mesh size smaller than the drug [207],
or by incorporating the drug or protein of interest into the backbone of the material or as a graft
23
[208]. Degradation controlled systems generally have potential for longer delivery than
diffusion mediated systems, where drug can diffuse through the polymer network; however,
degradation in vivo can be too slow and degradation products can have toxic side effects [19].
Micro- and nano-particles have been widely developed for drug delivery to slow the rate of
release for long time periods [19]. Particles can also be incorporated into polymer networks to
achieve extended delivery of molecules [209], and mechanisms for release include diffusion,
matrix or bulk degradation, and surface erosion [210].
Polymer systems can be crosslinked systems, physically interacting systems, or interpenetrating
networks. Hydrogels are crosslinked hydrophilic polymer networks that contain high water
content, and are formed by physical or chemical crosslinks. Physically-crosslinked hydrogels are
less stable than covalently crosslinked hydrogels, and generally have faster degradation rates.
Physical interactions include hydrogen bonding, hydrophobic junctions, and electrostatic charge
interactions [19], making these hydrogels ideal for minimally invasive injection techniques [10,
211] or environmentally responsive to temperature or pH changes [19, 212]. Chemically-
crosslinked hydrogels can be formed in situ through chemical techniques such as Michael-type
addition reactions [213], or through UV photopolymerization [214] at or just prior to the time of
implantation. However, chemical reactions such as these can cause proteins to be crosslinked
into the gel structure, and crosslinking and initiating components can also have toxic affects
[215]. Since hydrogel properties are easily controlled, they are excellent space-filling materials
for irregularly shaped gaps or cavities while minimizing the invasiveness of implantation.
Natural polymers, which include proteins and polysaccharides, are those that have biologic
origin and have the advantage of being similar to the natural biologic environment. Because of
this, they are often degraded by naturally present enzymes, and degradation products are
metabolized through physiological mechanisms [19]. However, since they are source specific,
they can produce a significant immunogenic response, although this can be modulated by
identification and modification of antigenic determinants [19]. Synthetic polymers can be less
immunogenic than natural polymers and are generally easier to chemically manipulate for
specific applications. The most common biomedical hydrogels include some derived from
natural sources, such as chitosan, collagen, hyaluronan, agarose, alginate, dextran, and fibrin; as
well as synthetic polymers, such as polyesters, poly(acrylic acid), and their derivatives like
poly(ethylene glycol) (PEG), poly(2-hydroethyl methacrylate) (PHEMA) and poly(N-isopropyl
24
acrylamide) (pNiPAAm). Poly(lactic acid) and polyglycolide are synthetical polyesters with
unique biomedical properties since lactic acid is produced within the body and therefore has no
immunogenicity [216]. PEG is a material that has been approved by the FDA for several
medical applications due to its biocompatibility and is used as coatings for nanoparticles or in
micelles to minimize immune detection. It has also been used to modify proteins to increase the
systemic half-life or biodistribution into tissues [217]. Clearly, the biomedical application for
which a material is needed will influence the choice of a natural or synthetic polymer. The
following sections describe in more detail the three polymers that were used for various aspects
of this research.
1.7.1 Hyaluronan
Hyaluronan (HA) is a naturally-derived polysaccharide composed of the repeating disaccharide
D-glucouronic acid β(1-3) N-acetyl-D-glucosamine. It is found in the extracellular matrix of
many tissues, including the CNS, and is a main constituent of synovial fluid [218]. The linear
chains range from 10-15,000 disaccharide units, and form random coils that are highly hydrated
and able to form entangled molecular networks [219]. At low strain frequencies, the molecules
are capable of aligning parallel to the direction of stress and flow, dissipating energy as viscous
flow and heat. At high strain frequencies, the molecules cannot adjust configurationally, and act
as elastic solids, which is why they are widely used as treatments for osteoarthritis and tissue
augmentation [220]. These shear thinning properties that allow the molecular chains to align
under force make HA an optimal material for an injectable gel because the molecular weight of
the polymer does not break down from the injection force.
Although concern surrounds natural polymers because of immunogenic affects, HA promotes
wound-healing by reducing inflammation and minimizing tissue adhesion and scar formation.
HA has differing properties dependent on its molecular weight where high and low molecular
weight references are generally thresholded at 3-6 x106 Da [220]. Low molecular weight HA
can stimulate angiogenesis [221], but can also increase the macrophage proliferation rate [222].
High molecular weight HA at concentrations greater than 1mg/ml decreases proliferation of and
increases apoptosis of macrophage in vitro [222]. Hyaluronan has found widespread use because
of these non-immunogenic and biocompatible properties [218].
25
Hyaluronan is degraded enzymatically by hyaluronidase, and more quickly in an inflammatory
environment [223]. Degradation can also be mediated by chain diffusion since there is little
interaction between the long MW chains. Chemical modification of HA can occur at the
carboxyl or hydroxyl groups, or with the amino group once the N-acetyl group is removed
(Figure 1.4) [220]. Chemical modification of HA can be accomplished using relatively simple
methods, and depending on the modification, can be used to modulate the properties of the
material. Unlike many other natural polymers, HA alone does not form a gel, however, this
facilitates blending it with other polymers. The shear thinning property of HA makes it optimal
for injectable applications, while the ease and diversity of modification sites allow drug delivery
properties to be tailored.
Figure 1.4: Chemical structure of hyaluronan (A), methylcellulose (B), and poly(ethylene
glycol) (C)
1.7.2 Methylcellulose
Methylcellulose (MC) is a carbohydrate that is derived from cellulose (Figure 1.4). When degree
of methylation is between 1.4 and 1.9, methylcellulose has inverse thermal gelling properties,
causing it to gel upon a temperature increase [224]. Water molecules form cagelike structures
around the hydrophobic moieties of MC at low temperatures. Upon a temperature increase, these
these cagelike structures break, exposing the hydrophobic methyl groups which then form
26
entropically favored hydrophobic junctions (Figure 1.5) [225]. In water, MC can form a weak
gel at physiological temperatures (37°C), but salt solutions decrease the gelation temperature,
thus increasing the stability of the gel at physiological temperatures [226]. This phenomenon
occurs because water molecules associate with salt molecules more readily than the hydrophobic
sugar units of MC, thus reducing its solubility. Although MC is not enzymatically degraded, MC
chains dissolve over time, thereby breaking the gel. Cellulose based membranes have been
shown to activate the complement pathway in the immune system [227-229], although this
response is lowered for modified cellulose structures [230, 231]. Despite the concern of
activating the immune system, MC was used as a scaffold for experimental traumatic brain
injury in vivo, and showed good biocompatibility over a two week time span [232, 233]. It has
also been used for peripheral nerve scaffolds and did not show adverse pathological reactions
over 8 weeks [232, 233]. These data support the use of MC in biomedical applications,
particularly for nervous tissue.
Figure 1.5: Gelation of MC caused by heating via formation of hydrophobic junctions
27
Blending HA and MC forms a viscous material at room temperature that forms a physically
crosslinked hydrogel when heated. When prepared in the salts that comprise the CSF, a 2% HA
and 7% MC formulation is a hydrogel at physiological temperatures [10]. The HA acts as a
thickener for the gel yet maintains its molecular weight upon injection since it is shear thinning.
These properties of HA make the solution more viscous, thus retaining the solution at the site of
injury during the gelation process. HA also has a “salting out” effect on MC, causing gelation to
occur more rapidly since the hydrophilic hydroxyls of HA draw water molecules away from the
hydrophobic MC chains. The MC component forms the physical crosslinks that produce the
hydrogel structure. These properties make this blend of HAMC is injectable through a small
needle (30 G) for use as a minimally invasive intrathecal delivery system for local administration
of drugs or proteins.
1.7.3 Poly(ethylene glycol)
Poly(ethylene glycol) is a hydrophilic, nonionic polyether with a narrow polydispersity. PEG is
unique in that it is capable of dissolving in both organic and aqueous solvents, and it can bind 2-
3 water molecules per ethylene oxide unit [234]. The binding of water and the flexibility of the
backbone chain have been postulated as reasons for the unique biomedical properties of PEG.
The process by which PEG is covalently conjugated to peptides or proteins is known as
PEGylation. This process has been used to improve molecular stability by increasing resistance
to proteolysis [235], and reducing immunogenicity by shielding immunogenic binding sites
[236]. PEG can decrease tissue clearance of proteins [237], thus enhancing its penetration into
tissue. It has been shown to increase the biologic half-life of several molecules and also enhance
distribution in tissues [238]. The distribution of PEGylated proteins in tissue depends on both
the physico-chemical factors of the PEG-protein conjugate and the physiological/anatomical
factors of the host tissue. Since PEG can be functionalized with a variety of reactive groups, it
can be attached to proteins using various chemistries, which are reviewed in [234]. In this work,
PEG was used to improve the diffusion of FGF2 into injured spinal cord tissue [239], as
discussed in detail in chapter 3.
28
1.8 Historical Context and Motivation for this Research Injectable polymers are excellent materials to utilize for localized drug delivery after SCI. The
intrathecal space is irregularly shaped, but is an optimal site for drug delivery since this route
circumvents the BSCB without causing damage to the cord. Additionally, a very small insertion
site is highly desirable to minimize a breach in the dura and leakage of CSF. Since the dura is
exposed after decompressive surgery, this offers a unique clinical window for this type of
treatment after SCI. Figure 1.6 illustrates an injectable system for intrathecal drug delivery,
which was first developed by Jimenez Hamann et. al. [9], consisting of a viscous collagen
solution that was injected into the intrathecal space. This collagen gel was used to co-deliver
recombinant human epidermal growth factor and FGF2 [87]. Compared to bolus injections, the
epidermal growth factor penetrated deep into the injured spinal cord and was found to be present
for up to 6 h within the cord. FGF2 was only detected on the outer periphery of the tissue, likely
because it is tightly regulated in vivo through extracellular matrix binding, thus inhibiting its
diffusion into the tissue. Nonetheless, these experiments verified that an injectable polymeric
drug delivery system could achieve the goals of localized delivery, and resulted in better tissue
penetration than a transient injection into the CSF of diffusible proteins. However, the collagen
gel was not entirely biocompatible for this purpose since it is a cell adhesive material, which led
to cellular build-up around the material.
Subsequently, Gupta et. al. [10] developed an injectable blend of hyaluronan and methylcellulose
for this purpose, improving upon the initial design of the collagen gel. The design criteria for
this material was: (1) Injectable through a 30 gauge needle, (2) Fast gelling to facilitate localized
drug delivery, (3) Biodegradable to avoid the need for removal, (4) Minimally swelling to not
further compress the injured cord, (5) Non-cell adhesive to prevent scar formation, and (6)
Biocompatible. In vitro work showed that HAMC met all of the first 5 design criteria [10], and
an in vivo study was done to test the biocompatibility. This biocompatibility/safety study was
done in both uninjured and injured rats to determine any potential impact this would have on
tissue recovery and motor function behavior. The in vivo results using the clip compression SCI
showed that compared to a control injection of artificial CSF intrathecallly, HAMC injected
intrathecally did not significantly increase the lesion volume or the number of apoptotic cells,
and in fact, showed a reduction in the inflammatory response following injury [10]. Behavioral
testing with the BBB locomotor scale showed animals that received HAMC performed better
29
than aCSF controls 1 week following injury and performed similarly from 2-4 weeks after injury.
Uninjured animals that received either aCSF or HAMC intrathecally did not show any loss of
function over the same period. These results suggested that HAMC was a safe and
biocompatible intrathecal drug delivery system.
The research described herein describes the research performed to further characterize HAMC as
a minimally invasive, intrathecal drug delivery system, as well as build on the existing
framework of HAMC to improve delivery of therapeutic agents. Initially, we studied the in vivo
degradation properties of HAMC to determine the longest period of delivery that could be
achieved with this promising drug delivery system. The neuroprotective molecule EPO, which
can cross the blood-brain barrier, was delivered with HAMC and compared to traditional routes
of delivery, specifically, an intraperitoneal (systemic) injection and intrathecal bolus injection.
This study was done to determine if localized delivery can improve upon these delivery routes
with respect to tissue and functional recovery, as well as to assess the potential of EPO as a
neuroprotective agent. Chapter 2 describes this work in detail, and has been published in Tissue
Engineering [240]. Following this work, PEG was conjugated to FGF2 to improve its diffusion
into spinal cord tissue, since previous work with the collagen gel showed that FGF2 did not
diffuse into tissue. Both FGF2 and PEG-FGF2 were delivered with HAMC in the intrathecal
space, and their diffusion pattern into injured spinal cord tissue was elucidated using ELISA
detection of serially digested tissue samples. This study was done to improve the delivery of
FGF2 to the injured spinal cord and also understand the diffusion characteristics of proteins
delivered from HAMC in the intrathecal space. Chapter 3 describes the results of this work, and
has been accepted for publication in the Journal of Controlled Release [239]. Finally, the
therapeutic benefit of FGF2 delivered with HAMC was studied with respect to the hemodynamic
changes that occur following SCI. Since the methods used to study hemodynamic properties
require large animal numbers, a new method utilizing dynamic CT was developed for this
purpose. Both this new method and the impacts of FGF2 delivered from HAMC are discussed in
a manuscript that is in review for submission to the Journal of Neurotrauma [241], and is
described in Chapter 4. The entirety of this dissertation describes the use of HAMC as an
intrathecal drug delivery system for neuroprotective agents after acute SCI.
30
Figure 1.6: Injectable polymeric drug delivery system for intrathecal delivery (copyright
Micheal Corrin).
©Michael Corrin
Injectable polymeric drug delivery system
31
2 Localized Delivery of Erythropoietin with the Hyaluronan-Methylcellulose Blend
A New Paradigm for Local and Sustained Release of Therapeutic Molecules to the
Injured Spinal Cord for Neuroprotection and Tissue Repair*
*This chapter has been peer-reviewed and published in Tissue Engineering with the following authors: Catherine Kang, Peter Poon, Charles Tator, and Molly Shoichet [240]
2.1 Abstract (Local Delivery of EPO)
After spinal cord injury (SCI), a complex cascade of events leads to tissue degeneration and a
penumbra of cell death. Neuroprotective molecules to limit tissue loss are promising; however
intravenous delivery is limited by the blood-spinal cord barrier and short systemic half-life.
Current local delivery strategies are flawed: bolus injection results in drug dispersion throughout
the intrathecal space and catheters / pumps are invasive and open to infection. Our laboratory
previously developed a hydrogel of hyaluronan (HA) and methylcellulose (MC) (HAMC) that,
when injected into the intrathecal space was safe and remarkably, had some therapeutic benefit
on its own. In order to test this new paradigm of local and sustained delivery, relative to
conventional delivery strategies, we tested for the first time the in vivo efficacy of HAMC as an
intrathecal drug delivery system by delivering a known neuroprotective molecule, erythropoietin
(EPO). In vitro studies showed that EPO was released from HAMC within 16 h, with 80%
bioactivity maintained. When the material alone was injected in vivo, individual fluorescent
labels on HA and MC showed that HA dissolved from the gel within 24 h, whereas the
hydrophobically associated MC persisted in the intrathecal space for 4-7 days. Using a clip
compression injury model of moderate severity, HAMC with EPO was injected in the intrathecal
space and, in order to better understand the potential of this delivery system, compared to the
therapeutic effect of both common delivery strategies – intrathecal EPO and intraperitoneal EPO
– and a control of intrathecal HAMC alone. Intrathecal delivery of EPO from HAMC resulted
in both reduced cavitation after SCI and a greater number of neurons relative to the other
delivery strategies. These data suggest that the localized and sustained release of EPO at the
tissue site by HAMC delivery enhances neuroprotection. This new system of intrathecal delivery
holds great promise for the safe, efficacious and local delivery of therapeutic molecules directly
to the spinal cord.
32
2.2 Introduction (Local Delivery of EPO) Treatment for the devastating condition of spinal cord injury (SCI) is currently limited, in part
due to the complexity of the pathophysiology after trauma and the lack of effective therapies
[242]. After an initial traumatic injury to the spinal cord, a cascade of events causes further
tissue damage. This process is termed the secondary injury and is characterized by ischemia,
hemorrhage, inflammation, and edema within the cord. These secondary events traumatize a
large area of tissue that was previously unaffected or sublethally affected by the primary insult,
and is thought to be responsible for the significant further loss of function that occurs after SCI
[28]. Recent research has focused on administering neuroprotective agents to interrupt this
cascade, thereby minimizing tissue degeneration.
High doses of therapeutic agents are required to cross the blood spinal cord barrier (BSCB) and
reach the site of injury when delivered by the traditional systemic route, often leading to
widespread side effects. Often only a brief period of delivery is achieved due to renal clearance
and the short half-life of molecules in blood. These limitations suggest that localized delivery of
these agents would result in greater neuroprotection and tissue sparing. Bolus delivery of
therapeutic agents to the intrathecal space has been investigated; however, the therapeutic
window is short due to drug clearance by cerebrospinal fluid (CSF) flow and absorption [243,
244]. Osmotic minipumps became popular in the last decade for localized and sustained
experimental delivery of drugs to the intrathecal space, and implanted pumps have been used in
humans for the delivery of analgesics and other agents [199]. However, widespread utilization
of pumps has been slow due to catheter blockage, infections, and formation of proliferative
lesions around the insertion site [115]. Our laboratory has pioneered an injectable, polymeric
drug delivery system for intrathecal drug administration whereby therapeutic agents dispersed
throughout a fast-gelling polymer can be released directly at the site of injury [9, 87]. A new
material blend of 2% hyaluronan (HA) and 7% methylcellulose (MC) was developed and shown
to be fast gelling within the intrathecal (IT) space, minimally invasive by injection through a
30G needle, non-cell adhesive, biodegradable, and biocompatible in vivo [10]. The goal herein
was to test the potential of this new paradigm of intrathecal HAMC delivery to conventional
intrathecal bolus and intraperitoneal delivery strategies.
33
Methylprednisolone, a steroid administered systemically, is one of the few agents that can cross
the blood-spinal cord barrier (BSCB) and shown to have a positive effect on recovery in clinical
trials, although there was only a mild benefit [245] and the trials have been widely criticized [6,
7]. Many other therapeutic agents are being investigated which show reduced apoptosis in
several cell types [161, 246], reduced demyelination [247, 248], reduced inflammation [249,
250], among others. These affects may directly prevent or indirectly lead to sparing of neurons
which can result in greater remodeling of the sensory and motor spinal tracts to improve
functional recovery post-injury. Recent reports show that the hematopoietic protein
erythropoietin (EPO) and EPO receptors, which play a role in central nervous system
development, are upregulated in the brain and spinal cord within minutes to hours after injury
[156, 157]. EPO has shown very positive functional recovery in stroke models, and the
neuroprotective benefit of this glycoprotein has been proven in hypoxic and ischemic neonatal
and adult rat brains [167-170]. Various groups have also shown a neuroprotective benefit of EPO
in spinal cord injury, both in ischemic and compressive/contusive models [49, 142, 157, 161].
The exact neuroprotective mechanism of EPO after SCI is still unclear, but it is postulated that
mechanisms may include reduction of ischemic damage [161], reduction of inflammation [162,
163], and sparing of white matter tracts in the spinal cord [49, 164]. EPO may also have more
direct actions on neurons, oligodendrocytes and astrocytes, all of which express receptors for this
molecule [156, 159, 165, 166]. Since EPO can cross the BSCB [172], it was an ideal molecule
to test in terms of delivery strategy in the context of neuroprotection after SCI. Since HAMC
forms a loose polymer network, fast drug release from this material was expected; however
Gorio et. al. [49] demonstrated that a single intraperitoneal (IP) injection of EPO provided
similar neuroprotective effects after SCI to repeated IP injections for several days, further
emphasizing EPO as a choice to test in intrathecal HAMC delivery.
The objective of this study was to investigate the benefit of intrathecal HAMC delivery relative
to conventional delivery strategies intrathecally and intraperitoneally. EPO was chosen as the
molecule to test in this new delivery paradigm because of its demonstrated neuroprotective
effects. We first studied the in vitro release of EPO from HAMC to better understand the time
course of release in vivo. The in vivo degradation of HAMC was also determined to ensure a
sufficient time window for in vivo EPO delivery. Finally, the in vivo tissue benefit of EPO
delivered in HAMC was compared to the conventional delivery strategies of intrathecal EPO
34
alone and intraperitoneal (IP) EPO alone, using intrathecal HAMC alone as a control for the
delivery vehicle. We hypothesized that the localized and sustained release achieved with
intrathecal delivery of EPO in HAMC would result in greater neuroprotection and tissue sparing
than the controls.
2.3 Materials & Methods (Local Delivery of EPO) Media and cells were purchased from ATCC (American Type Culture Collection, Rockville,
MD) and reagents were sterile filtered before use. Unless otherwise indicated all chemicals were
purchased from Sigma Aldrich Chemical Co. (Mississauga, ON) and used as received.
2.3.1 Material preparation
Sodium hyaluronate was purchased from Novamatrix (1.5x106 Da, Drammen, Norway) and
sterilized prior to use. To sterilize HA, a 0.1% HA solution in Millipore deionized water (dH2O)
was filtered through a 0.22 µm PES filter (Nalgene, Rochester, NY). The solution was then
lyophilized under sterile conditions by covering 50 ml tubes with nylon 0.22 µm filters
(Millipore, Billerica, MA), producing sterile HA powder. Methylcellulose was sterilized by
autoclaving for 20 min at 120°C. Artificial cerebrospinal fluid (aCSF) was prepared in dH2O
with 148 mM NaCl, 3 mM KCl, 0.8 mM MgCl2, 1.4 mM CaCl2, 1.5 mM Na2HPO4, and 0.2
mM NaH2PO4
Sterile HAMC was produced by mixing polymer solutions in a laminar flow hood. The MC
powder was added to half of the appropriate amount of sterile filtered aCSF at 90°C and vortexed
until all polymer particles were wetted. The remaining amount of aCSF was added cold at 4°C
and the solution was shaken on an ice bath for 30 min. HA powder was then added to the MC
solution, vortexed, and allowed to dissolve overnight at 4°C. This resulted in a 2% HA and 7%
MC solution.
[9].
2.3.2 In vitro release of EPO from HAMC
The EPO/HAMC blend was similarly prepared to that of HAMC, with EPO dissolved in cold
aCSF and added to MC that was dispersed in hot aCSF prior to the addition of HA. This order of
addition was important to achieve a well-mixed suspension of EPO in HAMC. The solution was
then loaded into a Hamilton syringe and 10 µl injected into the bottom of eppendorf tubes
35
containing 100 µl of aCSF, thereby approximating the large ratio of CSF to HAMC that is
expected in vivo by injection into the intrathecal space. These samples were incubated at 37°C
on an orbital shaker and aCSF was fully removed and replaced with fresh aCSF at t=0, 10 min,
20 min, 40 min, 1, 2, 4, 8, 16, 32, 64, and 128 h. An ELISA assay (R&D Systems, Quantikine
IVD EPO Kit) performed in triplicate was used to determine the concentration of EPO in the
aCSF that was removed at each time point.
2.3.3 Bioactivity of released EPO
A cellular assay was used to determine the activity of the released EPO on the basis of TF-1 cell
proliferation in response to EPO. Cells were maintained in RPMI-1640 media with 10% fetal
bovine serum, 1% penicillin/streptomycin, and 2ng/ml of granulocyte macrophage-colony
stimulating factor (GM-CSF). A 10 µl aliquot of HAMC containing EPO was injected into 90 µl
of culture media lacking GM-CSF. These samples were incubated for t=10 min, 20 min, 40 min,
1, 2, 4, 8, and 16 h, and the media was fully transferred to a fresh plate. A TF-1 cell suspension
was added to each sample to achieve a final cell density of 1x104 cells/ml in each well. A
standard curve was obtained using samples of EPO dissolved in media at known concentrations,
and kept in similar conditions to those with EPO in HAMC. All cells were incubated for 2 d at
37°C and 5% CO2
2.3.4 Degradation of HAMC in vitro and in vivo
, and cell density then assayed with Cell Titer 96, a substrate metabolized into
a colored product only by live cells. The absorbance was measured at 490 nm with a
VERSAmax tunable microplate reader and cell density was calculated. Based on the standard
curve correlating expected cell density to known concentrations of EPO, the concentration of
bioactive EPO released from HAMC was determined for samples collected at each time point.
Using carbodiimide chemistry HA was conjugated to a BODIPY-Fluorescein (BODIPY-FL)
hydrazide and MC was conjugated to Texas Red hydrazide for visualization within the
intrathecal space. For HA conjugation, a 1% HA in dH2O solution was adjusted to pH 4.6 and
1-ethyl-3-(3-diamethylaminopropyl) carbodiimide hydrochloride (EDC) was added at a 2:1
molar ratio to HA. N-hydroxysuccinimide (NHS) was added in a 1:1 molar ratio to EDC to
stabilize the intermediate and BODIPY-FL hydrazide was then added in a 1:4 molar ratio to HA
[251]. This was allowed to react at room temperature for 24 h, and was then dialyzed overnight
(12,000 MWCO, Spectrum laboratories Inc., Los Angeles, CA) and sterilized as described above
36
for unmodified HA. For MC, the hydroxyl groups were first modified to carboxyl groups with
bromoacetic acid using a protocol obtained from Hermanson [252]. Methylcellulose was added
to 1 M bromoacetic acid in 3 M NaOH and reacted at room temperature for 1.5 h. The reaction
was stopped by adding solid monobasic sodium phosphate and neutralizing with 6 N HCl and
fluorescent modification of MC was then performed as described above for HA, with Texas Red
hydrazide used in place of BODIPY-FL hydrazide. MC was then dialyzed overnight and
sterilized in the same way as unmodified MC. In the fluorescent HAMC solution, fluorescent
HA accounted for 0.5% of the total HA in the blend, and fluorescent MC accounted for 1% of
total MC in the blend, which was prepared by the same procedure as their unmodified
counterparts.
To determine if degradation of the labeled material matched that of the unlabeled HAMC, the
mass loss of each was determined over time in vitro. To maintain the ratio of HAMC to CSF
expected in vivo, 100 μl of unlabeled HAMC or fluorescent HAMC was injected into 1 ml of
aCSF which was equilibrated to 37 ºC. At t=0, 1, 2, 4, 8 h and 1, 2, 4, 7, and 14 d, aCSF was
fully removed from the samples and the remaining material was lyophilized and weighed. Mass
at each time was compared to that at t=0 to determine material loss over time.
Animal procedures were performed in accordance with the Guide to the Care and Use of
Experimental Animals developed by the Canadian Council on Animal Care and approved by the
Animal Care Committee at the Research Institute of the University Health Network. Fifteen
Sprague-Dawley rats (200-250 g; Charles River, Montreal, QC) were anesthetized by inhalation
of halothane, and a laminectomy performed at the T1-2 vertebral level. Fluorescently labeled
HAMC was injected intrathecally as described in Jimenez-Hamann et. al. [87]. Following
injection, the overlying muscles and fascia were sutured closed, and the rats were ventilated with
pure oxygen and placed under a heat lamp for recovery. Buprenorphine was administered every
12 h for 3 d post-surgery for pain management.
At t=0 (immediately after injection), 1, 2, 4, and 7 d, animals were administered a lethal dose of
sodium pentobarbital, and a 2 cm section of the spinal cord was removed at the T1-2 level.
Tissue was removed fresh to prevent the gel from dislodging during the fixation process. Cords
were sectioned parasagittally and imaged on a Leica DMRB inverted microscope with Stereo
Investigator Software, version 6. For each component of HAMC, the fluorescent area within a
37
given intensity range was calculated for each image. Fluorescence loss observed over time was
indicative of bulk hydrogel degradation, which was calculated according to equation (1) where
Ao is the initial fluorescent area (at t=0) within intensity limits and At
is the fluorescent area at
time t within the same intensity limits:
(1)
The loss of fluorescence was scaled according to the % of HA or MC that was labeled. For HA a
scaling factor of 4 was used because 1/4 of the total HA was fluorescent-HA; for MC a scaling
factor of 7 was used because 1/7 of the total MC was fluorescent-MC.
2.3.5 EPO Efficacy Study
The operative procedure for 47 Sprague-Dawley rats (220-300g; Charles River, Montreal, QC)
was performed as described in the degradation study, except that rats sustained an SCI before
injection. SCI was performed with the aid of an operating microscope by placing a modified
aneurysm clip calibrated to a closing force of 35 g on the spinal cord. The cord was acutely
compressed with this clip for 60 sec, as previously described [43]. Immediately following the
injury, one of four injections (10 µl) was performed: 1) injection of HAMC containing EPO into
the intrathecal space (n=10); 2) HAMC without EPO injection into the intrathecal space (n=13);
3) bolus injection of EPO into the intrathecal space (n=12); 4) bolus injection of EPO into the
intraperitoneal cavity (n=12). An injection volume of 10 µl was constant amongst all groups,
and all EPO solutions were prepared at 100 IU/µl. Animal weights varied from 250-300 g,
resulting in a dosage of 3500±200 IU/kg. This dosage was used based on previous studies with
EPO in SCI where an intraperitoneal dosage of 5000 IU/kg produced the greatest functional
benefit [49]. Previous work in our lab showed that intrathecal injection of aCSF was comparable
to HAMC injection, and therefore the aCSF group was excluded from this study [10]. Post-
surgical care of rats was identical to that in the degradation study except that rat bladders were
manually expressed 3 times daily.
38
2.3.6 Functional Assessment
To determine the efficacy of EPO in enhancing motor behavior after injury, open field motor
function was assessed using the Basso, Beattie, and Bresnahan (BBB) scoring method [84] daily
for 1 week and then weekly thereafter for 6 weeks. Each hindlimb was ranked by two blinded
observers and concurrently videotaped. BBB scores range from 0 (no hindlimb movement) to 21
(normal gait behavior) and are used to assess functional improvement after injury and treatment.
2.3.7 Histology & Immunohistochemistry
Animals for the degradation study were sacrificed at t=0 (immediately following injury), 1, 2, 3,
and 5 d following intraperitoneal injection of an overdose of sodium pentobarbital. Fresh tissue
was harvested to prevent dislodging HAMC from the intrathecal space, then cryoprocessed and
stored at -80°C until cut into 20 µm parasagittal sections. Images were obtained at 10x with a
Leica DMRB inverted microscope and Stereo Investigator Software (version 6). The area of HA
and MC at each timepoint was determined by pixel intensity using Image J software.
Animals for the EPO efficacy study were sacrificed 42 d after surgery and perfused intracardially
with 4% paraformaldehyde under deep anesthesia. A 2 cm segment of spinal cord encompassing
the injury site was harvested from each animal and cryoprocessed. Cords from three to four
animals from each group were sectioned parasagittally at 20 µm thickness, and every sixth
section was stained with either 1) Luxol Fast Blue/Hematoxylin and eosin (LFB/H&E) for
general morphology and to assess cavity area (volume and area measurements were equivalent
for the 35g clip injury model that we have utilized here [51]), 2) ED-1 (activated macrophage
stain), or 3) glial fibrillary acidic protein (GFAP – activated astrocyte stain). Images were taken
at 5x (LFB/H&E) or 10x (ED-1, GFAP) with a Leica DMRB inverted microscope using Stereo
Investigator Software (version 6), and sections in the maximal cavity area were analyzed for area
of macrophages and astrocytic scar formation based on pixel intensity using Image J software.
Cords from three other animals in each group were cross-sectioned at 20 µm for 1 cm caudal to
the injury, and every fourth section stained with NeuN (Neural cell body stain) or CC-1
(Oligodendrocyte stain) to assess sparing of these cell types caudal to the injury. Images were
obtained at 20x with a Olympus BX61 Microscope with Image Pro Plus Software (version 5.1).
Neurons were counted manually and oligodendrocytes were counted in Image Pro Plus, which
counted only objects within limits on intensity and object size.
39
2.3.8 Statistical Analysis
All statistics were performed using Microsoft Excel. One way ANOVA followed by the Tukey’s
post hoc t-test was used to compare BBB scores of all groups. A student’s t-test was used to
compare the maximal cavity area and neuron counts. Differences were accepted to be
statistically significant at p < 0.05. All errors are given as standard deviations.
2.4 Results (Local Delivery of EPO)
2.4.1 In vitro release of EPO from HAMC
Prior to studying the therapeutic and tissue benefit of intrathecal EPO delivered in HAMC
relative to conventional delivery techniques, we investigated the in vitro release profile where the
volume ratio of EPO/HAMC injected into aCSF mimicked that of the animal model. The release
profile of EPO was investigated over 128 h using an ELISA assay, from which it was determined
that 99% of EPO was released from HAMC during the first 16 h (Figure 2.1A). Peppas and
others [253] have previously established models showing that the amount of drug release is
proportional to tn where n is indicative of the transport mechanism from polymeric drug delivery
systems. In unidimensional diffusion mediated release from a slab geometry, n is equal to 0.5
and a linear relationship exists between drug release and t0.5
Figure 2.1
for ~60-80% of release [253]. This
is a good approximation for the HAMC hydrogel geometry both in vitro in a test tube where the
gel is a flat slab and in vivo in the intrathecal space where the polymer blend is bound by the
spinal cord and the dura. In B, cumulative release in vitro is plotted versus t0.5 and a
linear relationship is maintained for the first ~75% of release (R2
A TF-1 cellular assay demonstrated that 80% of the released EPO remained bioactive over the 16
h release time (
= 0.992), providing evidence
that EPO release from HAMC is diffusion mediated. Deviation from the model was expected for
>80% release due to depletion of drug within the hydrogel, resulting in n > 1 [254].
Figure 2.1C). The ratio of bioactive EPO released to cumulative EPO released
over time remained constant, showing that EPO bioactivity did not decrease with time. The 20%
loss of activity was likely a result of the increased temperature needed to dissolve MC and obtain
a good dispersion of EPO within HAMC. With most of the EPO released being bioactive, the
therapeutic benefit of local intrathecal delivery was investigated.
40
Figure 2.1: (A) EPO release from HAMC over 128 h shows fast release within 16 h. (B)
Cumulative release of EPO vs. the square root of time demonstrates a diffusion-mediated
release profile for the first 75% of EPO released. (C) EPO released from HAMC is 80%
bioactive over time relative to cumulative EPO released. (=Cumulative EPO released;
=Bioactive portion of EPO released)
41
2.4.2 Degradation of HAMC in vitro and in vivo
Having established in vitro EPO release and bioactivity profiles from HAMC, we were interested
in understanding the HAMC degradation profile in vivo. An in vitro degradation profile, shown
in Figure 2.2A, demonstrated that fluorescently labeled HAMC had similar mass loss over time
as that of the unlabeled material. Figure 2.2B shows longitudinal spinal cord sections where
each of the HA and MC are labeled with a different fluorescent molecule. The profile of HAMC
loss was followed over time by monitoring the change in area of fluorescence within the
intrathecal space. HA degraded quickly, exhibiting a ~95% loss in fluorescent area after 24 h.
In contrast, MC showed an initial degradation of ~65% after 24 h and then continued to persist
within the intrathecal space for at least 4 days. After 7 days, traces of neither HA nor MC could
be detected (Figure 2.2C).
Figure 2.2 (see next page): (A) In vitro mass degradation is the same for labeled and
unlabeled HAMC. (B) Brightfield image of a parasagittal spinal cord section with
corresponding fluorescent images of HA (Green) and MC (Red) in the IT space
immediately after injection. (C) Loss in the area of fluorescence shows that HA degrades
faster than MC in the IT space. Scale bars =1mm. (n=3, mean ± standard deviation
plotted) ((=mass loss of labeled HAMC; =mass loss of unlabeled HAMC)
=fluorescence loss of HA; =fluorescence loss of MC)
42
43
2.4.3 In vivo delivery of EPO with HAMC
To determine the efficacy of intrathecal HAMC delivery relative to conventional intrathecal
bolus and intraperitoneal delivery, EPO was chosen as the therapeutic molecule due to its known
neuroprotective effects and ability to cross the BSCB, and then further compared to a control of
intrathecal HAMC alone. After SCI, secondary events that lead to inflammation and extensive
cell death result in fluid accumulation or edema within the spinal cord [255]. Subsequently, a
cyst or a cavity forms within the cord, filled with cellular debris and toxic factors [3]. The
ultimate size of this fluid-filled cavity is proportional to the severity of the injury that is
sustained, but the size may be reduced due to tissue sparing. To quantify any change in the size
of the cavity due to HAMC or EPO, the maximal cavity area was measured in parasaggital
histological sections stained with LFB/H&E (Figure 2.3A). A reduction in the cavity area
suggests tissue sparing, which is important for both neuroprotective and regenerative strategies.
Figure 2.3B shows that the maximal cavity area for EPO delivered intrathecally from HAMC
was significantly lower than that of the HAMC control (p<0.05), demonstrating the benefit of
intrathecal HAMC delivery of EPO over HAMC alone. While not significantly different,
intrathecal HAMC delivery of EPO had less tissue loss than either intrathecal EPO or
intraperitoneal EPO delivery. Thus, of the three delivery strategies, intrathecal HAMC delivery
of EPO resulted in the greatest tissue sparing, demonstrating the benefit of local and sustained
release.
44
Figure 2.3: (A) Parasaggital section of a representative spinal cord of an animal that
received IT administration of EPO-loaded HAMC at maximal cavitation stained with
LFB/H&E. (B) Comparison of maximal cavity area measured from parasaggital spinal
cord sections. (n=4, mean + standard deviation; horizontal bars show statistical significant
based on a t-test, *p<0.05)
The general morphology of the tissue suggested that some cells may have been spared due to
local delivery of EPO to the spinal cord. To determine which cells were spared, both neurons
and oligodendrocytes were counted. Significantly higher numbers of neurons were counted in
animals that received intrathecal HAMC delivery of EPO than animals that received EPO by
either conventional route of delivery – intrathecal bolus or intraperitoneal (Figure 2.4) – thereby
showing an added tissue benefit of localized and sustained delivery with HAMC. Interestingly,
45
bolus intrathecal injection shows similar neuron counts to intraperitoneal injection, substantiating
studies that show EPO can cross the BSCB, particularly after an injury to the CNS [172].
Intrathecal HAMC alone resulted in highly variable neuron counts, demonstrating that its
previously reported benefit [10] alone is insufficient to achieve neuroprotection.
Figure 2.4: Neuron counts observed in the ventral horns caudal to the injury site based on
NeuN staining. (n=3, mean + standard deviation; horizontal bars show statistical
significance based on t=test, *p<0.05)
Based on previous reports that describe white matter sparing with EPO delivery [49], we
postulated that there may also be a reduction in demyelination within the spinal cord, translating
into a greater number of oligodendrocytes. However, oligodendrocyte counts showed no
significant differences between any groups (data not shown). While it is generally known that
reactive glial cells secrete factors that can lead to cell death, this action is more commonly
associated with astrocytes than oligodendrocytes. Since no significant increase in
oligodendrocytes or decrease in glial scar formation from reactive astrocytes (GFAP
immunoreactivity, data not shown) was observed, it is unlikely that either of these indirect
mechanisms of neuroprotection occurred. Combined with the neural sparing observed, these
results suggest a more direct action of EPO binding to EPO receptors on neurons.
EPO has also been shown to decrease inflammation [163], and for this reason ED-1
immunoreactivity was investigated (Figure 2.5). ED-1 stains activated macrophages and
46
microglia within the spinal cord and the area over which these cells reside can be measured by
pixel intensity of the images obtained. At the maximal cavity area, the total area of inflammation
was measured, but EPO delivery did not appear to reduce the inflammatory area. Since it is
possible for EPO delivered intraperitoneally to cross the BSCB, this small amount could
decrease the macrophage response similar to EPO delivered intrathecally. Also, because
inflammation was measured 6 weeks post-injury, the difference in the inflammatory response
may have abated by this stage of recovery. Previous work with HAMC alone showed a
significant decrease in inflammation relative to injection of aCSF without any therapeutic agents
[10], which was attributed to the wound healing property of HA [219].
Figure 2.5: Comparison of maximal inflammation area measured with ED-1
immunostaining of macrophages and microglia from parasaggital spinal cord sections.
(n=3, mean ± standard deviation).
The functional benefit of EPO delivered with HAMC was tested by locomotor functional
analysis using Basso, Beattie Bresnahan scoring. Interestingly, at 7 days after EPO injection,
animals that received intrathecal HAMC delivery of EPO behaved significantly better than those
that received intrathecal bolus EPO (p<0.05); however, in subsequent weeks, these animals
behaved similarly to all other animals. Although Gorio et al. [17] showed a functional benefit
with intraperitoneal EPO injections, this study was later found to be irreproducible [256]; the
present study also showed no functional benefit from EPO delivery, either intrathecal or
intraperitoneal except at 7 days from HAMC (Figure 2.6). The BBB scores were very similar for
47
all groups, including those that received the control injection of HAMC alone. Within 3 weeks
of injury, most animals had reached a plateau in functional recovery, with all animals exhibiting
weight support of hindquarters.
Figure 2.6: Open field motor scores as assessed by the BBB scale over six weeks. (*shows
statistical significance measured by ANOVA; n=9 to 11, data shown as mean ± standard
deviation) (=IT administration of EPO-loaded HAMC; =IT administration of HAMC
alone; =IT administration of EPO; =IP administration of EPO)
2.5 Discussion (Local Delivery of EPO)
The diffusional release observed in vitro can be interpreted using a scaling argument where drug
release for a characteristic time (τ) follows equation (2):
(2)
where L is the characteristic length and Ddrug is the diffusion coefficient of the drug [257]. The
thickness of the gel was estimated at 2 mm based on the volume of gel and dimensions of the
reservoir. The diffusion coefficient of the 30 kDa EPO molecule was estimated to be on the order
of 1x10-10 m2/s [258]. Thus, the characteristic diffusion time for EPO from HAMC is on the
order of 10 h, consistent with the 16 h release observed experimentally. While this time scale is
relatively short in the scheme of tissue remodeling that can occur for weeks or months after SCI,
48
intrathecal HAMC delivery of EPO likely has longer lasting effects on the tissue because the
apparent concentration of EPO at the site of injury remains more than 100 times longer after
injury when delivered with HAMC than either intrathecal bolus or intraperitoneal delivery of
EPO, based on the average velocity of CSF flow [12].
Upon injection in vivo, the temperature of HAMC increases from room temperature to 37°C,
resulting in the formation of hydrophobic junctions of MC. These junctions act as physical
crosslinks within the hydrogel, preventing fast dissolution of MC. However, an initial loss of
MC is observed, with a slower degradation observed over a period of days. The initial 65% loss
of MC in the first 24 h after injection may occur due to either quick dissolution prior to full
gelation of MC or polydispersity of the MC chains, where lower molecular weight chains erode
more quickly than higher molecular weight chains. Furthermore, non-homogeneous distribution
of methyl groups on MC could result in chains with only a few hydrophobic domains dissolving
away more quickly, leaving only chains stabilized by several hydrophobic junctions.
The degradation of HAMC observed in vivo was faster than that observed in vitro, where
degradation occurred over a period of 10-15 days. The difference may be due to continual CSF
flow in the intrathecal space, which amplifies the hydrogel dissolution rate. Also, the spatial
constraints on the gel in vivo are greater than those in vitro; the spinal cord and dura limit gel
layer thickness in vivo and thinner slabs degrade more quickly due to increased disentanglement
of polymer chains [254]. Importantly, the degradation of the fluorescently labeled HAMC was
not statistically different from that of the unlabeled HAMC (Figure 2.2A), validating the in vivo
degradation results.
The rapid loss of HA from the hydrogel blend can be explained either by dissolution or
degradation. Since HA forms crosslinks with neither MC nor itself, it is expected to dissolve
away from the gel, assisted by CSF flow in the intrathecal space. The natural enzyme
hyaluronidase has a high specific activity for cleaving the N-acetyl-D-glucosamine in HA, and at
concentrations found in serum [259], it would degrade all of HA in HAMC within minutes of
injection. Because this rate would cause complete loss of HA before tissue harvesting at initial
timepoints and hyaluronan is present in the CSF [260], it is unlikely that enzymatic degradation
contributed significantly to the loss of HA in the intrathecal space. This suggests that dissolution
is likely the dominant mechanism for HA loss after injection into the intrathecal space.
49
The concentration of EPO at the injury site is likely lowest for EPO delivered intraperitoneally
due to dilution in the blood, presence of the BSCB and CSF flow. Moreover, EPO causes red
blood cell production [261] and thus systemic administration can lead to thickening of the blood
and increase the time needed to clear the hemorrhage after injury. These may account for the
lack of neuroprotection after intraperitoneal relative to intrathecal HAMC delivery of EPO.
Similarly, intrathecal bolus delivery of EPO resulted in lower neuroprotection relative to
intrathecal HAMC delivery of EPO, likely due to the fast clearance by CSF flow. The
improvement in neuron counts may be attributed to longer residence time and bioavailability of
EPO released from HAMC at the site of injury combined with the wound healing properties of
HA. Spatial limitations of the hydrogel-spinal cord interface also cause a constant supply of
EPO on the surface of the spinal cord, and EPO diffusing from HAMC into the cord in this
manner may provide enhanced neuronal protection due to drug-receptor binding on the surface
of neurons where EPO receptors are present [155, 159]. Cultured primary neurons treated with
EPO have been protected from excitotoxic cell death [155], and experiments with EPO in stroke
and traumatic brain injury models [170] have also shown this neuroprotective effect in hypoxic
and ischemic environments [169]. While EPO delivery in HAMC showed greater
neuroprotection than HAMC alone, the difference was not statistically significant; the high
standard deviation in the HAMC group suggests that HAMC may play some role in
neuroprotection, but not strong enough to reproducibly achieve neuron sparing. Together these
data suggest a synergistic effect between the drug delivery system and the delivered molecule.
We have shown here that EPO delivered locally with HAMC can increase neuronal sparing as
compared to the traditional routes of drug delivery. The enhanced tissue and neuron sparing
achieved demonstrate the importance of our localized release strategy to limit degeneration and
ultimately as a means to enhance functional recovery.
Behavioral recovery was significantly greater at 1 week post-injury for animals that received
intrathecal HAMC delivery of EPO relative to animals that received intrathecal delivery of EPO
alone; however, animals that received the bolus intrathecal delivery of EPO alone had acute
respiratory complications during surgery. Within 15-30 s of injection, animals that received an
intrathecal bolus injection of EPO alone began gasping deeply for breath, and after 3-5 gasps
became apnoeic. Of the 16 rats undergoing surgery in this group, 12 exhibited this behavior, 4 of
which could not be resuscitated and died. The resuscitative efforts consisted of terminating
50
anesthesia and applying external compression of the thoracic region until a pulse was identified
and respiration restored. A recent report shows that slowed breathing is observed when EPO is
administered in the central nervous system [262]; however cessation of breathing was not listed
as a side effect. This unexpected respiratory complication may have contributed to the lack of
early functional recovery in this group. This was not observed in animals that received
intraperitoneal EPO or intrathecal HAMC alone, but was observed for intrathecal HAMC
delivery of EPO in 2 of the 10 animals, both of which were successfully resuscitated.
Interestingly, when the dose of the intrathecal bolus EPO was reduced to 80%, to adjust for the
bioactivity of EPO in HAMC, gasping was observed, but animals did not cease to breathe,
suggesting that this effect is dose-dependent. Although the enhanced locomotor functional
benefit of EPO delivered via HAMC was only apparent at 7 days post-injury, delivery with
HAMC allowed slower release of EPO and fewer respiratory complications relative to intrathecal
infusion of EPO alone. The functional benefit observed at 7 days may have occurred due to
other secondary injury events such as reduced inflammation or decreased apoptosis at early time
points; EPO has been shown to reduce the inflammatory response following traumatic brain
injury [163, 263, 264] and decrease apoptosis [163, 264]. The early improvement suggests a
localized delivery strategy with a longer release profile would sustain recovery further. This
illustrates an added benefit of HAMC in terms of improved safety for intrathecal drug
administration and the potential for even greater functional recovery with prolonged delivery
times.
The rate of dissolution of HAMC in vivo and the release rate of EPO in vitro suggests that this is
a useful strategy for the localized release of neuroprotective molecules with an early therapeutic
window. This work showed that intrathecal HAMC delivery of EPO provided the greatest
neuroprotective benefit in terms of tissue sparing after SCI. This was exemplified by decreased
cavitation over HAMC alone and a greater number of spared neurons as compared to the
intrathecal bolus of EPO and intraperitoneal EPO. Additionally, respiratory complications
associated with EPO administration were minimized when delivered with HAMC. In ongoing
studies, the intrathecal HAMC delivery system is being further refined for prolonged degradation
and release of other therapeutically relevant molecules with a focus on improving
neuroprotection and regeneration after spinal cord injury.
51
3 Enhancing Localized Delivery of Fibroblast Growth Factor 2
Poly(ethylene glycol) Modification Enhances Penetration of Fibroblast Growth Factor 2
in Injured Spinal Cord Tissue from an Intrathecal Delivery System*
*This chapter has been peer-reviewed and is in press in the Journal of Controlled Release with the following authors: Catherine Kang, Charles Tator, and Molly Shoichet [239].
3.1 Abstract (Penetration of FGF2) Spinal cord injury is a condition for which there is no effective treatment and clinical drug
delivery techniques for central nervous system conditions are limited by the blood-spinal cord
barrier. Our lab has developed an injectable drug delivery system consisting of a biopolymer
blend of hyaluronan and methylcellulose (HAMC) that can sustain drug release for up to 24 h in
the intrathecal space. Fibroblast growth factor 2 (FGF2) has great potential for treatment of
spinal cord injury due to its angiogenic and trophic effects, but previous studies showed no
penetration into spinal cord tissue when delivered locally. Conjugation to poly(ethylene glycol)
(PEG) is known to improve penetration of proteins into tissue by reducing clearance and
providing immunogenic shielding. We investigated conjugation of PEG to FGF2 and compared
its distribution relative to unmodified FGF2 in injured spinal cord tissue when delivered
intrathecally from HAMC. Importantly, PEG conjugation nearly doubled the concentration of
FGF2 in the injured spinal cord when delivered locally and, contrary to previous reports, we
show that some FGF2 penetrated into the injured spinal cord using a more sensitive detection
technique. Our results suggest that PEGylation of FGF2 enhanced tissue penetration by reducing
its rate of elimination.
3.2 Introduction (Penetration of FGF2) Spinal cord injury is a serious condition affecting nearly 300,000 people in the US. Most
frequently, a spinal cord injury is caused by broken fragments of the vertebral column
compressing the spinal cord causing immediate and delayed cell death and blood vessel rupture.
The primary injury causes limited tissue damage, but the series of events in the hours to days
after injury include hemorrhage, ischemia, hypoxia, inflammation, and edema, all of which
52
contribute to the characteristic tissue degeneration and major loss of function [265]. The current
clinical treatment for this condition includes an intravenous injection of high dose
methylprednisolone; however, the clinical benefit of this treatment has been highly debated [6, 7]
and at present there are no other proven treatments [242].
Drug delivery to the central nervous system (CNS) is particularly difficult due to the blood-
spinal cord barrier and the dura and arachnoid membranes that surround the cord, both of which
are effective in preventing the passage of most drug and protein therapeutics delivered
systemically. This emphasizes the need for other methods than the traditional intravenous and
oral delivery strategies, which often require high doses for penetration and can lead to significant
and undesirable side effects. Intrathecal drug delivery strategies to circumvent these barriers
have been developed, such as bolus delivery, implantable catheters, or sustained delivery from
minipumps. However, these strategies are invasive and prone to infections [115] and bolus
intrathecal delivery is evanescent. To achieve sustained intrathecal delivery, our lab has
developed a minimally invasive, injectable drug delivery system consisting of a physical blend of
hyaluronan and methylcellulose (HAMC) which has proven safe for in vivo use [10]. A
therapeutic agent can be dispersed either alone or in polymeric nanospheres within this hydrogel
[209], and the hydrogel alone has shown promise in vivo to provide short-term, localized
delivery of therapeutic molecules when injected into the intrathecal cavity following spinal cord
injury [240]. To relieve pressure on the spinal cord, decompressive surgery is often required to
remove vertebral fragments. During this procedure, HAMC could be injected into the intrathecal
space directly over the injury site (Figure 3.1A) to achieve maximal therapeutic benefit from
localized release.
Fibroblast growth factor 2 (FGF2) is a protein with multiple modes of action in the CNS and has
been tested both in experimental models of spinal cord injury [138, 139] and clinically for the
treatment of stroke [193]. Aside from its well known trophic properties in promoting
angiogenesis [180-182], FGF2 has also been shown to be neuroprotective [187, 188]. It
stimulates endogenous ependymal cells to proliferate when co-delivered with EGF [87] and can
reduce permeability of the blood-brain barrier [190, 191]. Additionally, FGF2 has enhanced
functional recovery in rats after spinal cord injury when delivered locally from an osmotic
minipump [138, 139]. However, because FGF2 does not cross the blood-spinal cord barrier, a
localized delivery system is necessary and this need is underscored by FGF2 being mitogenic
53
and upregulated in cancer [182] reflecting the limitations of systemic administration [197].
Notwithstanding the interesting functional benefits ascribed to FGF2, its penetration into the
spinal cord when delivered locally beneath the dura and arachnoid mater has only been studied
once. This study was based on immunohistochemistry and demonstrated that FGF2 delivered
intrathecally from a collagen hydrogel did not penetrate into spinal cord tissue [87]. To enhance
tissue penetration and increase local concentrations, we conjugated poly(ethylene glycol) (PEG)
to FGF2, using a methodology similar to PEG modification of other proteins [266, 267]. We
then investigated the penetration and distribution of PEG-FGF2 in the spinal cord relative to
unmodified FGF2.
PEG modification has been used to improve molecular stability [234], as well as reduce
immunogenicity [236] and decrease tissue clearance [237], thus enhancing its penetration into
tissue. It has been shown to increase the biologic half-life of several molecules and also enhance
distribution in tissues [238]. The distribution of PEGylated proteins in tissue depends on both
the physico-chemical factors of the PEG-protein conjugate and the physiological/anatomical
factors of the host tissue. To better understand FGF2 biodistribution in spinal cord tissue, we
developed a tissue Enzyme-Linked Immunosorbent Assay (ELISA) technique to study the
penetration of FGF2 vs. PEG-FGF2 delivered locally from the injectable HAMC hydrogel. This
technique was more sensitive than the previously used immunohistochemical method and
demonstrated tissue penetration of both FGF2 and PEG-FGF2. Our results suggest that the
greater tissue penetration observed for PEG-FGF2 than FGF2 is ascribed to reduced elimination.
3.3 Materials & Methods (Penetration of FGF2)
All chemicals were purchased from Sigma-Aldrich Chemical Co (Mississauga, ON) and used as
received unless otherwise noted. Media and cells were purchased from ATCC (American Type
Culture Collection, Rockville, MD) and reagents were sterile-filtered in a 0.22 µm Stericup filter
before use.
3.3.1 Synthesis of PEG-FGF2
Poly(ethylene glycol) was modified with a maleimide functional group prior to conjugation with
FGF2. PEG monomethyl ether (PEG, 5000 Da) was dissolved in dichloromethane and reacted
54
with a 3x molar ratio of succinimidyl 4-(N-maleimidomethyl) cyclohexane-1-carboxylate
(SMCC, Pierce, Chicago, IL) in the presence of 1.5x molar ratio of triethylamine under nitrogen.
After 24 h PEG-maleimide (PEG-mal) was precipitated dropwise in cold ether and filtered
through a 0.42 µm glass microfiber filter. PEG-mal was dried under vacuum overnight to
remove excess ether and stored at -20ºC. Matrix-assisted laser desorption/ionization mass
spectrometry (MALDI-MS, Voyager Elite Mass Spectrometer, Toronto Integrated Proteomics
Lab) was performed to characterize PEG (CH3(OCH2CH2)113OH (M+) 5,000 Da, found 5,088
Da) and PEG-mal (CH3(OCH2CH2)113OC11H14NO4
Recombinant human fibroblast growth factor 2 (FGF2, Biovision, Mountain View, CA) was
dissolved in 5mM Tris buffer to a final concentration of 1 mg/ml. PEG-mal was added in a 5x
molar ratio to 300 µl of FGF2 and was reacted at room temperature on a shaker at 500 rpm for 3
h. A heparin affinity binding column (GE Healthcare Biosciences AB, Uppsala, Sweden) was
used to remove free PEG from the reaction mixture. Purified PEG-FGF2 was dialyzed (10,000
MWCO) overnight at 4ºC and subsequently lyophilized in conical tubes with sterile nylon 0.2
µm filter caps prior to use.
(M+) 5,236 Da, found 5,223 Da).
Sodium Dodecyl Sulfate – Polyacrylamide Gel Electrophoresis (SDS-PAGE) was performed
with a 12% acrylamide resolving gel and 5% stacking gel (Biorad, Constant Voltage = 120V) to
confirm the PEG-FGF2 product had formed and that free PEG was fully removed. A Benchmark
pre-stained protein ladder (Invitrogen, Burlington, ON) was used for molecular weight
classification. Proteins were stained with Simply BlueTM
3.3.2 Bioactivity of PEGylated FGF2
Safestain (Invitrogen, Burlington, ON)
and PEG was stained for 20 min with sequential submersion in 2.5% barium chloride and 1.3%
iodine/2% potassium iodide using a modified protocol from [268].
Dose dependent proliferation of a Balb/3T3 fibroblast cell line (ATCC, CCL-163) was used to
investigate the bioactivity of FGF2 and PEG-FGF2 products. Cells were maintained in
Dulbecco’s Modified Eagle’s Medium supplemented with 1% Penicillin/Streptomycin and 10%
calf bovine serum. To test the activity of PEG-FGF2 relative to that of FGF2, Balb/3T3 cells
were cultured in serum free media with either FGF2 or PEG-FGF2 mitogens at 10, 20, and 40
ng/ml with a 20:1 weight ratio of Heparin:FGF2. These cells were incubated for 24 h at 37°C
and 5% CO2, and cell density then assayed with Cell Titer 96, a substrate metabolized into a
55
colored product only by live cells. The absorbance was measured at 490 nm with a VERSAmax
tunable microplate reader and cell density was calculated relative to a standard curve of known
cell density.
3.3.3 In vitro release of FGF2 vs. PEG-FGF2 from HAMC
Sodium hyaluronate (Novamatrix, 1.5x106 Da, Drammen, Norway) was sterilized by filtering a
0.1% solution through a 0.2 μm filter and lyophilizing prior to use. Methylcellulose (13x103
Da), was sterilized similarly. Following lyophilization in conical tubes with sterile nylon 0.2 μm
filter caps, sterile HAMC was produced by mixing polymer solutions in a laminar flow hood.
Artificial cerebrospinal fluid (aCSF) was prepared in dH2O with 148 mM NaCl, 3 mM KCl, 0.8
mM MgCl2, 1.4 mM CaCl2, 1.5 mM Na2HPO4, and 0.2 mM NaH2PO4
3.3.4 In vivo distribution of FGF2 in spinal cord tissue
[9]. The MC and HA
powders were sequentially dissolved in a solution of 100 μg/ml of either FGF2 or PEG-FGF2 in
aCSF at 4°C. This resulted in 100 μg/ml FGF2 or PEG-FGF2 loaded into a 2% HA and 7% MC
solution. HAMC with no protein was similarly prepared as a control to ensure HAMC did not
interfere with the ELISA assays. A 100 μl aliquot of this solution was injected into the bottom
of centrifuge tubes containing 900 µl of aCSF at 37ºC, approximating the ratio of HAMC to CSF
that is expected in vivo by injection into the intrathecal space of a rat. These samples were
incubated at 37°C on an orbital shaker and aCSF was fully removed and replaced with fresh
aCSF at t=1, 3, 6, 12, 24, and 48 h. A sandwich ELISA assay (R&D Systems, Human FGF basic
Duoset) was used to determine the concentration of FGF2 or PEG-FGF2 in the aCSF that was
removed at each time point (n = 4).
All animal procedures were performed in accordance with the Guide to the Care and Use of
Experimental Animals developed by the Canadian Council on Animal Care and approved by the
Animal Care Committee at the Research Institute of the University Health Network. Fifty two
adult Sprague-Dawley rats (200-250 g; Charles River, Montreal, QC) were anesthetized by
inhalation of halothane, and a laminectomy performed at the T1-2 vertebral level. A modified
aneurysm clip calibrated to a closing force of 26 g was applied to the spinal cord at T1 for 1 min
to simulate a spinal cord injury as previously described [43]. A durotomy was performed
immediately caudal to the injury site with a 30 G bent beveled needle and then a 30 G blunt, bent
tipped needle was inserted into the intrathecal space (see Figure 3.1A) for injection of HAMC.
56
To allow for HAMC gelation, the needle was held in the intrathecal space for 1 min before
removal, at which time a small plug of HAMC would form at the durotomy site and prevent
leakage of CSF [10]. Animals received a 10 μl injection in the intrathecal cavity of HAMC
loaded with either 10 µg/ml FGF2 (n=5/timepoint) or 10 µg/ml PEG-FGF2 (n=5/timepoint).
To control for cross-reactivity in the ELISA of endogenous rat FGF2 with the recombinant
human FGF2 injected, animals were injured but not injected (n=3/timepoint). Following
injection, the overlying muscles and fascia were sutured closed; rats were ventilated with pure
oxygen and placed under a heat lamp for recovery. Buprenorphine was administered post-
surgery for pain management.
At t=0 (immediately after injection), 3, 6, and 24 h, animals were administered a lethal dose of
sodium pentobarbital, and a 2 cm section of the spinal cord was removed at the T1-2 level.
Removal of fresh tissue prevented loss of HAMC in the intrathecal space and FGF2 from tissue
during the fixation process. Tissue was flash frozen in 2-methylbutane on ice and subsequently
cut parasagittally on a tissue chopper (McIlwain, Redding, CA) into 1 mm thick slices. The most
medial two slices were then cut into 300 μm longitudinal sections dorsoventrally so that FGF2
could be detected spatially through the depth of the spinal cord (See Figure 3.1B). Homogenate
buffer was prepared in distilled H2O with 20 mM HEPES, 10 mM KCl, 1.5mM MgCl2, 1mM
ethylenediamine tetraacetic acid (EDTA), 1mM ethylene glycol tetraacetic acid, 1mM
dithiothreitol, 1 mM phenylmethanesulphonyl fluoride. Each 1mm x 300 µm x 2 cm segment of
spinal cord was digested in homogenate buffer using a Pellet Pestle (Kontes, USA). Samples
were centrifuged for 15 min at 15,000 RPM at 4ºC and the supernatant collected. A sandwich
ELISA assay (R&D Systems, Human FGF basic Duoset) performed in duplicate on supernatant
samples was used to determine the concentration of FGF2 and PEG-FGF2 in each tissue
segment. Low concentrations of rat FGF2 cross reacted with human FGF2 antibodies; these
concentrations were then subtracted from concentrations detected in the cords with rhFGF2 and
PEG-rhFGF2 that was delivered locally for each timepoint, respectively.
57
Figure 3.1: (A) Diagram of injection paradigm shows hydrogel spreading following
intrathecal injection. Arrows indicate potential directions for protein flux from HAMC
into the CSF or into spinal cord tissue. (B) Tissue sampling paradigm for observing depth
of dorsoventral penetration. Spinal cord segments of 2 cm spanning the entire rostro-
caudal spread of the hydrogel were removed and then sectioned sagittally into 1mm thick
sections. The most medial two sections were then sliced longitudinally into 300 μm
58
sections. These 1mm x 300µm x 2cm sections were used to measure the dorsoventral
penetration of FGF2 or PEG-FGF2 from the dorsal aspect to the center of the spinal cord.
3.3.5 Statistical Analysis
One way analysis of variance (ANOVA) followed by Tukey’s post-hoc t-test was used to
compare the depth of dorsoventral penetration concentrations of FGF2 and PEG-FGF2.
Differences were considered statistically significant at p<0.05. All error bars shown represent
standard deviations.
3.4 Results (Penetration of FGF2)
3.4.1 PEGylation of FGF2
A maleimide functionalized PEG (Figure 3.2) was conjugated to free FGF2-cysteine thiols. The
crystal structure of FGF2 reveals four free cysteine residues that do not participate in disulfide
bonding in the active FGF2 structure or in the well known binding sites for heparin sulfate
proteoglycans [178]. These free cysteines also do not interfere with the binding sites for the FGF
receptors 1 and 2 [269].
Figure 3.2: FGF2 with free
cysteine groups reacts with
PEG-mal to produce single
PEGylated FGF2 and double
PEGylated FGF2 (with some
unreacted, non-PEGylated
FGF2).
Gel electrophoresis showed 3
bands in the purified PEG-
FGF2 mixture (Figure 3.3, lane
3): ~17kDa, ~25 kDa, and ~35
kDa. Although there are 4 free cysteines on FGF2, Cys 96 and Cys 78 are more readily available
on the surface of the protein for conjugation [176]. Cys 34 and Cys 101 face the internal portion
59
of the protein when the active conformation is maintained, making the conjugation of PEG-
maleimide difficult on these residues [176]. For this reason, the PEG/FGF2 reaction mixture
delivers 3 distinct bands: FGF2, mono-PEG-FGF2, and di-PEG-FGF2. The molecular weights
expected were ~22kDa and ~27 kDa for single and double PEGylation respectively; however,
bands were observed shifted higher at ~25 and ~35kDa. This is likely due to the complexation
of SDS with PEG, which is known to slow the movement of PEGylated proteins in SDS-PAGE
[270].
When the reaction was allowed to proceed for several hours longer, 5 bands were observed (data
not shown), also shifted higher than the expected molecular weights, with the higher molecular
weights corresponding to three and four PEG molecules bound to FGF2. Because these products
with 3 and 4 PEGs would likely exhibit reduced bioactivity due to the conformational change
necessary to expose the internal Cys residues, the protocol which produced only single, double,
and unPEGylated FGF2 was used, and this mixed product is henceforth referred to as PEG-
FGF2.
Lane 3 in Figure 3.3 was loaded with products previously separated on a heparin affinity binding
column and shows no PEG band at low molecular weights. Lane 4 shows the flow through of
the reaction solution, with a dark band of PEG observed near ~6 kDa. These results show that
free PEG was fully removed from the PEG-FGF2 solution prior to dialysis. Densitometry of the
bands showed that of the products, 18% was non-PEGylated FGF2, 44% was mono-PEG-FGF2,
and 38% di-PEG-FGF2. Thus 82% of the FGF2 was
PEGylated.
Figure 3.3: Representative gel electrophoresis shows
molecular weight ladder, control FGF2, products of PEG-
FGF2 reaction isolated from heparin binding column, and
flow through PEG from heparin binding column.
60
3.4.2 PEG-FGF2 Bioactivity
Dose dependent proliferation of Balb/3T3 cells was used to ensure that the PEG-FGF2
maintained bioactivity after PEG modification and processing steps. The cell density increased
similarly with increasing concentrations of either FGF2 or PEG-FGF2 (Figure 3.4), indicating
that PEG-FGF2 maintained dose dependent activity following the PEGylation and purification
steps.
Figure 3.4: Dose dependent activity of FGF2 (white bars) and PEG-FGF2 (black bars)
show that bioactivity based on cell density is maintained after PEGylation and subsequent
purification. (n=5, mean ± standard deviation are shown).
3.4.3 FGF2 Release from HAMC In vitro
The in vitro release profiles of PEG-FGF2 and FGF2 from HAMC were compared to understand
whether PEG modification would influence diffusion of FGF2 and to establish the duration of
release. Since the HAMC hydrogel is comprised of 9% solids and 91% buffer and is highly
porous, a diffusive mechanism of release was expected based on previous results with other
proteins of comparable size [240]. The release profiles of PEG-FGF2 and FGF2 were similar
and release was complete after 24 h (Figure 3.5).
61
Figure 3.5: In vitro release of FGF2 (□) and PEG-FGF2 () from HAMC drug delivery
system shows diffusive release within 24 h in almost identical concentrations. (n=5, mean ±
standard deviation are shown).
3.4.4 In vivo Distribution of FGF2 and PEG-FGF2
Previously, our laboratory found that FGF2 did not penetrate into spinal cord tissue based on
detection by immunohistochemistry where the limit of detection was 100 µg/ml [87]. This
restricted delivery of FGF2 motivated the present study to enhance the diffusive penetration into
the spinal cord of FGF2 by PEG modification. To overcome the limited sensitivity of
immunohistochemistry, we used the more sensitive tissue ELISA technique that increased the
sensitivity by four orders of magnitude to 10 ng/ml. With ELISA, both FGF2 and PEG-FGF2
were detected in spinal cord tissue when delivered locally (Figure 3.6). The spinal cord was
removed and frozen with the dura intact to enable measurement of the concentration of FGF2
and PEG-FGF2 in both the drug delivery system and the tissue. At 3 h, protein release from
HAMC was constant since the concentration in the gel did not change from t = 0 h. At this time,
a significantly greater concentration of PEG-FGF2 than FGF2 was detected in the spinal cord
tissue (Figure 3.6B). There was generally a greater concentration of PEG-FGF2 than FGF2 at all
depths of penetration from the dorsal surface of the spinal cord, with significant differences at
0.6 and 1.75 mm from the dorsal surface (Figure 3.6B). Concentrations within the spinal cord
62
were highest near the dorsal surface in close contact with HAMC; however, because HAMC
spread around the spinal cord tissue, diffusion of PEG-FGF2 and FGF2 was likely from both
dorsal and ventral sides, thereby accounting for the increase in protein concentration in the center
of the cord. Additionally, cerebrospinal fluid transport of the protein through the central canal
may have contributed to these elevated protein concentrations. At 6 h, the concentration in the
gel had dropped to ~1 µg/ml and release was no longer constant due to drug depletion.
Concentration profiles at 6 h show that concentrations of PEG-FGF2 tended to be higher than
FGF2 throughout the cord, but not statistically significant at this time (Figure 3.6C). At 24 h, no
FGF2 or PEG-FGF2 could be detected in the gel, and only small amounts (10-15 ng/ml) were
detected in the spinal cord (data not shown). We used Equation 1, based on concentrations
detected in tissue, to better understand FGF2 and PEG-FGF2 diffusion in vivo [271, 272]:
(1)
where C is the concentration in the tissue, Co is the initial concentration in the polymer delivery
system, x is the distance of the protein from the polymer and a is the characteristic thickness of
the polymer gel that is bound by the spinal cord and the dura. The φ is termed the
diffusion/elimination modulus [271, 273] and is defined as:
(2)
where k is a lumped elimination parameter accounting for cellular interactions and D is the
diffusion coefficient. Because this equation is applicable when drug delivery is at steady state,
estimated concentration profiles were obtained for t = 3 h only, where drug release was constant.
In vivo steady state model predictions using equation 1 above, as has previously been described
[273, 274] for similar systems, were fit to the data for up to 1.2 mm into the cord. These
predicted concentration profiles of FGF2 and PEG-FGF2 are shown in Figure 3.6B, from which
the φ values were calculated using the least fit squares method. This model fits those points
closest to the surface of the cord up to ~1.2 mm into the tissue, where release is most similar to a
point source. Deviation occurs beyond this depth likely due to concentric protein flux from other
surfaces of the cord (see Figure 3.1B, inset) as well as from CSF transport of protein through the
63
central canal. Importantly, this analysis provides useful insights into mechanism that accounts
for the greater tissue penetration of PEG-FGF2 vs. FGF2 by providing an understanding of the
relationship between the elimination constant (k) and diffusion coefficient (D). The φ values,
based on in vivo concentration values of and , were 1.2 and 0.7, respectively.
Because a is constant, the φ values indicate the relationship between elimination and diffusivity
of FGF2 and PEG-FGF2, where indicates greater elimination than diffusivity.
Figure 3.6: Cumulative comparison shows
a greater amount of PEG-FGF2 () in the
injured spinal cord than FGF2 (□) when
delivered from the intrathecal HAMC
drug delivery system. B At 3 h post-
intrathecal injection, higher
concentrations of PEG-FGF2 () than
FGF2 (□) have diffused into the cord.
Lines indicate the predicted concentration
profiles of PEG-FGF2 (solid) and FGF2
(dotted). C At 6 h post- intrathecal
injection, drug depletion from the
hydrogel results in a plateau of drug in the
tissue, but elevated concentrations of
PEG-FGF2 () are still observed in the
spinal cord. Asterisks indicate statistical
significance at p<0.05. (n=5, mean ±
standard deviation are shown).
64
3.5 Discussion (Penetration of FGF2) Intrathecal injection of HAMC containing either FGF2 or PEG-FGF2 allows local release of the
protein at the site of spinal cord injury. The HAMC hydrogel forms a gel spatially bound by the
dura and arachnoid membrane and the spinal cord from which the protein can diffuse laterally
into the CSF, radially into the spinal cord tissue, as well as across the dura and arachnoid
membrane into the epidural space (Figure 3.1A). Since the dura and arachnoid membrane
contributes to the blood-spinal cord barrier and is highly impermeable to proteins, minimal
diffusion is expected in this direction [275]. The spinal cord-hydrogel interface and the CSF-
hydrogel interface are the main sites for drug release from the hydrogel. The gel spreads over
the surface of the spinal cord and fills much of the intrathecal space at the site of injection,
thereby resulting in a significant spinal cord-hydrogel interface; however, protein flux into the
CSF is also expected. Penetrability of proteins into the spinal cord tissue is affected by diffusion
of the molecule in the extracellular space (which is bounded by cell membranes), as well as by
the elimination rate dictated by surface receptor binding or internalization of the protein,
degradation, and blood-spinal cord barrier clearance. Penetrability increases with dispersivity
and decreases with elimination, as is described in equations 1 and 2 above.
To further probe how PEG affected penetrability, we used the φ values calculated from in vivo
concentration values of FGF2 and PEG-FGF2, (1.2 and 0.7, respectively), which showed an
overall decrease in φ due to PEGylation. While the absolute values of k and D cannot be
determined for FGF2 and PEG-FGF2 within tissue, the ratios of each indicates relative
differences between the two proteins, which can be used to better understand why greater
concentrations of PEG-FGF2 are observed in the spinal cord. This ratio is shown as:
(3)
The ratio of obtained from the in vivo concentration profiles yields a ratio of
1.7. Given that FGF2 has a higher diffusion/elimination modulus, we then investigated how the
relative differences of the elimination and diffusion constants, which define φ, were affected by
PEGylation.
65
Absolute diffusion coefficients for FGF2 and PEG-FGF2 in tissue cannot be calculated due to
the complexities of protein-tissue interaction, but the relative difference of diffusion coefficients
between FGF2 and PEG-FGF2 can be estimated based on the ratio of molecular weight, where
[276]. The weighted average MW of PEG-FGF2 based on the relative amounts of
the reaction products was 23.2 kDa. The ratio of diffusion coefficients,
, indicated that diffusion should decrease due to PEGylation. Because
D is inversely proportional to φ, if penetration of each molecule was only affected by diffusion
then φ would be expected to increase as a result of PEGylation, and result in
. However, in vivo measurements yield ,
suggesting that PEGylation impacted the elimination parameter. The relation of k to φ (Equation
2) suggests that elimination would have to decrease significantly after PEGylation, with more
than a three-fold change needed to produce the nearly two-fold change in φ that was observed,
where . Since k is a lumped elimination parameter, it can be impacted
by several biological factors. Elimination can occur through intra- or extra-cellular degradation
of the protein, phagocytosis by cellular components, surface receptor binding, and clearance via
the bloodstream. PEG is known to enhance molecular stability of proteins [234], and
conjugation to FGF2 could decrease elimination by reducing extracellular degradation of the
protein. In addition, PEG conjugation is known to mask proteins via preventing phagocytosis by
the reticulo-endothelial system, which has been most commonly studied in renal clearance of
intravenously delivered PEG conjugates [217]. In the CNS, the reticulo-endothelial system is
represented by specialized immune cells called microglia, which become activated macrophages
following an injury. These cells phagocytose foreign material and cellular debris at the injury
site. If exogenously delivered FGF2 is detected as foreign, it will be cleared from the
extracellular space as it diffuses into the tissue. In Figure 3.6B, concentrations of FGF2 are
highest near the surface of the spinal cord where the drug loaded HAMC is releasing protein into
the tissue. As FGF2 diffuses into the tissue, macrophages that accumulate at the injury site may
internalize the foreign recombinant human FGF2. The PEG-FGF2 is thought to evade
phagocytosis due to the PEG chains that shield the exogenous FGF2, thus allowing greater
66
concentrations to reach the center of the cord. The analysis described here suggests that
PEGylation of FGF2 reduces elimination, further supporting previous reports of PEGylation
leading to enhanced molecular stability and biological evasion of macrophage phagocytosis.
3.6 Conclusions (Penetration of FGF2)
Using a reproducible method for PEGylation of FGF2, the penetration of FGF2 and PEG-FGF2
were compared: FGF2 and PEG-FGF2 were detected at ng/ml concentrations within the injured
spinal cord. These results demonstrate that exogenous FGF2 penetrates into the spinal cord
when delivered locally, and importantly, that higher concentrations of PEG-FGF2 vs. FGF2 can
be achieved. Analysis of the diffusion/elimination modulus suggests that elimination of FGF2
decreased with PEGylation, supporting biological models of PEG-mediated penetration of
proteins in vivo. These results support the concept that therapeutic proteins can be locally
delivered to the injured spinal cord with a minimally invasive injection of HAMC into the
intrathecal space.
67
4 Live Animal Imaging of Post-Injury Spinal Cord Blood Flow
Spinal Cord Blood Flow and Permeability Measured by Dynamic Computed
Tomography Imaging in Live Animals After Localized Delivery of Fibroblast Growth
Factor 2*
*This chapter is in preparation for submission to the Journal of Neurotrauma (2010) with the following
authors: Catherine Kang, Richard Clarkson, Charles Tator, Ivan Yeung, and Molly Shoichet.
4.1 Abstract (Hemodynamics with FGF2) Following spinal cord injury, profound vascular changes lead to ischemia and hypoxia of spinal
cord tissue. Since fibroblast growth factor 2 (FGF2) has angiogenic effects, its delivery to the
injured spinal cord may attenuate the tissue damage associated with ischemia. To limit systemic
mitogenic effects, FGF2 was delivered to the spinal cord from a gel of hyaluronan and
methylcellulose (HAMC) injected in the intrathecal space and compared to controls of HAMC
alone and artificial cerebrospinal fluid alone. Dynamic perfusion computed tomography (CT)
was employed for the first time in small animals to serially measure blood flow and permeability
in the injured and uninjured spinal cord. Spinal cord blood flow (SCBF) and permeability-
surface area (PS) measurements were obtained near the injury epicenter and at two regions
rostral to the epicenter. As predicted, SCBF measurements decreased and PS increased after
injury. FGF2 delivered from HAMC after injury restored SCBF towards pre-injury values in all
regions, and increased blood flow rates at 7 d post-injury compared to pre-injury measurements.
PS was stabilized at regions rostral to the epicenter of injury when FGF2 was delivered with
HAMC, with significantly lower values than aCSF controls at 7 d in the region farthest from the
epicenter. These data demonstrate that both localized delivery of FGF2 improves spinal cord
hemodynamics following injury and perfusion CT is an important technique to serially measure
these parameters in small animal models of spinal cord injury.
68
4.2 Introduction (Hemodynamics with FGF2) Traumatic spinal cord injury (SCI) is a serious condition that often causes paralysis and severe
loss of function. A recent study shows that more than 1.2 million people in the US report being
paralyzed due to SCI, and nearly 50% of those people reported “a lot of difficulty in movement”
or “a complete inability to move” [5]. Pharmacotherapy for acute SCI is currently limited to
intravenous administration of the steroid methylprednisolone; however its efficacy in improving
functional recovery has been questioned [6]. The early secondary events following the initial
spinal cord trauma lead to rapid and profound tissue changes, including hemorrhage,
inflammation, edema, and ischemia, all of which have significant impacts on tissue and
functional recovery. Several drug treatments using systemic delivery are designed to target and
limit these secondary injury mechanisms and have undergone clinical trials, as reviewed in [114,
242]; however, none has shown sufficient improvement in human functional recovery to be
established as a standard clinical practice. Re-establishing blood flow to the injured tissue could
have a neuroprotective effect by limiting degeneration resulting from ischemia.
Fibroblast growth factor 2 (FGF2) has been shown to both promote angiogenesis [180-182] and
be neuroprotective [187, 188]. Ex vivo studies suggest that FGF2 reduces permeability of the
damaged blood-brain barrier [190, 191]. In vivo studies demonstrate that FGF2 enhances
functional recovery in spinal cord injured rats when delivered locally from an osmotic minipump
[138, 139]. Thus, we hypothesized that FGF2 can improve spinal cord blood flow (SCBF) and
reduce permeability following injury, yet requires a local delivery strategy because it does not
cross the blood-spinal cord barrier (BSCB) [182]. While systemic drug delivery is the most
common practice clinically, it requires that molecules cross the BSCB, and high doses are
necessary to reach therapeutic levels at the site of injury, often leading to undesirable side
effects. Drug delivery systems have been designed to circumvent these limitations, such as bolus
intrathecal delivery and implantable catheters/minipump devices. However, these systems are
either transient or can cause complications, such as infection [115]. To overcome these
limitations, our lab has developed a minimally invasive, intrathecal drug delivery system that is
safe for local delivery of therapeutic agents to the site of injury [10, 240]. This localized
delivery system, composed of a biopolymer blend of hyaluronan and methylcellulose (HAMC),
is injected into the intrathecal space and locally releases FGF2 into the injured spinal cord for at
least 6 h [239].
69
In order to evaluate the potential benefit of local FGF2 delivery from HAMC to the injured
spinal cord, a dynamic method to serially study hemodynamics in the same animal over time was
desired. Early work in cerebral blood flow using gas clearance techniques were well
characterized and modeled by Kety and Schmidt based on Fick’s principles of diffusion [58, 59].
Utilizing inert gases to which the BSCB is completely permeable, the intravascular space, i, and
the extravascular space, e, are considered to be a single compartment and rate constants for gas
movement are equal between the extravascular and the intravascular spaces (kei and kie in Figure
1A). In this method, an inhalable and inert gas is first allowed to permeate the tissue until
saturation, and the clearance of the gas is then measured by electrodes placed within the tissue as
a function of time to determine the flow rate. SCBF is then calculated as shown in equation (1)
[59]:
(1)
where QT
Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) now have improved
resolution, and when combined with dynamic contrast enhancement blood flow and permeability
can be measured after contrast agents are injected into the bloodstream [66-74]. These
techniques build on the models used previously, but employ different assumptions and indicator
techniques. With MRI and CT, gadolinium-chelates and iodinated contrast agents, respectively,
(t) is the quantity of gas in tissue T at time t, W is the weight of the tissue, and the
denominator is the difference between the arterial and venous curves for gas concentration over
time. Although very useful for dynamically measuring blood flow, tissue damage occurs from
the probes that are inserted in the tissue for measurement. Other methods for measuring SCBF,
such as autoradiography, require a timed infusion of a radioisotope and subsequent sectioning of
the tissue [277]. The radioisotope is detected in tissue sections and quantified using
densitometry to assess regional flow into grey and white matter in the spinal cord. Indicator
fractionation techniques are similar, but instead use labeled microparticles in place of a
radioisotope [277]. More recently, the Laser Doppler technique has been used, but since it
requires the probe to be on the surface of the spinal cord, it is limited to either a relatively short
timeframe after the animal has undergone the injury or a terminal time point [278].
70
can be introduced into the bloodstream and the perfusion rate measured as they pass through the
tissues. Because these contrast agents or tracers do not diffuse freely in the tissue as gases
would, the intravascular and extravascular spaces within the tissue can be viewed separately (see
Figure 1B), unlike models developed by Kety and Schmidt. A consensus paper by Tofts et. al.
[75] describes this model in detail, known as the Modified Tofts Model, including the limitations
under different conditions and the association of fitting parameters to physiological parameters.
In this two-compartment model, the tissue is composed of the intravascular and extravascular
components, and a permeable tracer moves from the intravascular space to the extravascular
space at a volume rate constant of Ktrans. Because the tracer is sequestered in the plasma of
blood, the rate constant kep is defined as the rate from the extravascular space to the plasma
portion of blood. The unidirectional transport of tracer is dependent on the permeability and the
surface area of the endothelial layer that separates the intravascular and extravascular spaces in
the selected region. Thus, this method calculates the product of permeability and surface area of
endothelium in selected regions, known as the permeability-surface area product, or PS. The
generalized kinetic model for two compartments as previously described by Tofts et. al. [75] is
shown in equation (2a):
(2a)
and with the initial conditions that Cp = CT = 0 at t=0, this equation simplifies to:
(2b)
where Cp is the concentration in the plasma (defined as Ca/(1-Hct) where Hct is the blood
hematocrit), and Ce is the concentration in the extravascular space (defined as CT/ve, where ve is
the fractional volume of the extravascular space). This model can be used for any tissue, but the
parameters will vary based on the assumptions applied to the system. In a flow-limited situation
where a tracer is very permeable across the endothelial layer (e.g., where PS >> F), the equation
further simplifies to the Kety model shown in equation (3) where contrast enhancement is related
to flow (F) in the following manner:
71
(3)
where CA = Cp(1-Hct) and CV= (1-Hct)Ce (tissue density ρ is required since F is per unit mass
of tissue). SCBF measurements made with the models developed by Kety have customarily been
reported as ml/min/100g of tissue rather than per g of tissue, and this convention is maintained
here to allow for comparison with previous work. In a permeability-limited situation, where the
tracer is not very permeable to the endothelial layer (e.g., where F >> PS) then the situation
simplifies to equation (4):
(4)
where the units of PS are ml/min/g. Thus, in a flow limited situation, Ktrans = Fρ(1-Hct) (PS >>
F), and in a permeability limited situation, Ktrans = PSρ (F >> PS). In the central nervous
system, the BSCB is mostly impermeable to many compounds, thus F >> PS and the Ktrans
A simpler method for measuring F has been estimated from the first pass of an iodinated
compound through the tissue, and flow is only dependent on the maximum amount flowing into
the tissue shown in equation (5) [68, 279, 280]:
value
can be used to estimate PS. Thus, Equation 4 was utilized in these studies to calculate PS values
in the uninjured and injured spinal cord.
(5)
where flow is equivalent to the ratio of the maximal slope of the tissue intensity curve and the
peak height of the arterial curve. In this case, the perfusion rate through tissue is only depicting
the perfusion of the intravascular component, ignoring the extravascular component. The
underlying assumption is that in the short time used to measure peak tissue enhancement (usually
done in the first 10 sec), an insignificant amount of the tracer has permeated into the tissue and
thus can be ignored. This equation is only valid for cases where the tracer is not extremely
permeable to the membrane, and thus is measuring tracer uptake into the tissue rather than tracer
72
washout. Therefore, equation 5 was used to determine SCBF in the uninjured and injured spinal
cord from the first pass of contrast through the tissue.
The dynamic perfusion CT method has been used in several applications in both animal and
human studies. Initially, the method was used to investigate renal blood flow [68, 279, 280], but
further technical developments allowed measurements of tumor blood flow in cerebral tumors
[67, 73], lung cancers [72, 74], as well as renal and hepatic metastases [68, 70, 71]. A recent
report discusses the first use of CT perfusion for SCBF in humans [66], but this technique has
not yet been reported for small animals. In the present paper we evaluate dynamic CT imaging
for repeated measurements of SCBF and PS in rats after SCI, and with this technique assess the
effect of local delivery of FGF2 in HAMC on SCBF and permeability of the injured spinal cord.
A B
Figure 4.1: (A) One-compartment and (B) two-compartment model of blood flow.
4.3 Materials & Methods (Hemodynamics with FGF2)
All chemicals were purchased from Sigma-Aldrich Chemical Co (Mississauga, ON) and used as
received unless otherwise noted.
73
4.3.1 HAMC preparation
2 wt% HA and 7 wt% MC was prepared as previously described [239]. Briefly, sterile
methylcellulose (13x103 Da) and sodium hyaluronate (Novamatrix, 1.5x106 Da, Drammen,
Norway) powders were sequentially dissolved into artificial cerebrospinal fluid (aCSF) in a
laminar flow hood. The aCSF was prepared in dH2O with 148 mM NaCl, 3 mM KCl, 0.8 mM
MgCl2, 1.4 mM CaCl2, 1.5 mM Na2HPO4, and 0.2 mM NaH2PO4
[9]. To distribute
recombinant human FGF2 (Biovision, Mountain View, CA) in HAMC, a solution of 10 μg/ml of
FGF2 dissolved in aCSF was used in place of aCSF alone in the procedure above. This resulted
in 10 μg/ml FGF2 loaded into a 2% HA and 7% MCsolution.
4.3.2 In vivo Surgical Procedures
All animal procedures were performed in accordance with the Guide to the Care and Use of
Experimental Animals developed by the Canadian Council on Animal Care and approved by the
Animal Care Committee at the Research Institute of University Health Network. Twenty-four
adult Sprague-Dawley rats (200-250 g; Charles River, Montreal, QC) purchased with implanted
jugular vein catheters were anesthetized by inhalation of isoflurane, and a laminectomy
performed at the T1-2 vertebral level. Animals were imaged with dynamic perfusion CT
immediately following the laminectomy to obtain normal, pre-injury measurements. A modified
aneurysm clip calibrated to a closing force of 26 g was applied to the spinal cord at T2 for 1 min
to produce SCI as previously described [43]. A durotomy was performed immediately caudal to
the injury site with a 30 g bent beveled needle and then a 30 g blunt, bent tipped needle was
inserted through the durotomy into the intrathecal space. Animals received either 10 µl of aCSF
(n=8), 10 µl of HAMC (n=8), or 10 µl of FGF2 loaded HAMC (10 µg/ml FGF2, n=8) through
the needle that was then maintained in the intrathecal space for 1 min before removal to allow
HAMC gelation in the intrathecal space [10]. Following injection, the overlying muscles and
fascia were sutured, rats were ventilated with pure oxygen, and then placed under a heat lamp for
recovery. Buprenorphine was administered post-surgery for pain management. CT imaging was
performed at this time to obtain immediate post-injury measurements, and serial imaging was
performed at 1 d post-injury and 7 d post-injury, in the same animals.
74
4.3.3 Computed Tomography Imaging Acquisition with Dynamic Contrast Enhancement
Animals were kept under isoflurane anaesthesia for the duration of the CT scans, which were
obtained with a MicroCT scanner (Locus Ultra MicroCT, GE Medical Systems; Milwaukee, WI)
in the Spatio-Temporal Targeting and Amplification of Radiation Response Facility (Toronto,
ON). A scout x-ray was performed to define the level of interest corresponding to approximately
C5 through T4. A perfusion scan was then performed over the selected region, at an in-plane
resolution of 0.15 mm x 0.15 mm with a slice thickness of 0.15 mm covering a range of z = 13.5
mm at an energy of 60 kV and 90 mA. The iodinated contrast agent VisipaqueTM
(270 mg I/mL,
GE Healthcare, Milwaukee, WI) was injected as the scan started by means of a foot-pedal
controlled injection pump (NE-1000 Single Syringe Programmable Pump, New Era Pump
Systems, Inc., Farmingdale, NY). The contrast was warmed to 37⁰C to reduce viscosity and
delivered in the first 10 s of the scan at a rate of 124.2 mL/hr (total 354 μL delivered) through the
implanted jugular catheter. A 3 minute multiphase scan protocol was used, acquiring a volume
every second for the first 30 s, then every 10 s for the remaining 150 s.
4.3.4 Image Analysis
Post-processing was done using the open source ClearCanvas Software (Toronto, ON) with a
Dynamic Contrast Enhancement (DCE) tool developed at the University Health Network in
Toronto. A region of interest in the ascending aorta superior to the heart (at T3) was defined and
the time-intensity curve produced for the arterial input function. A tissue region of interest for
the spinal cord was then defined using average spinal cord measurements for long and short axes
of the spinal cord at each vertebral level [281] and using the remaining portions of the vertebrae
as boundaries (see Figure 2). Tissue-intensity curves were produced for 3 regions: 1) a region
immediately adjacent to the injury epicenter between T1 and T2 which is immediately rostral to
T2; 2) a region centered at the vertebral body of T1; and 3) a region between T1 and C7. Each
region corresponded to a 150 µm thick segment of tissue. An average hematocrit of 0.4 was
used for all measurements [281]. The onset time at which contrast was seen to enter each region
was manually determined and the tissue-intensity data was fit to a curve using the DCE tool.
75
Values for Ktrans
, initial slope in the first 10 s, and the peaks of the arterial and tissue curves were
calculated using the Modified Tofts Model, as described above. This procedure was performed
for each animal, at each time that a CT scan was obtained (t = pre-injury, 1 h post-injury, 1 day
post-injury, and 7 d post-injury).
4.3.5 Statistical analysis
All statistics were performed using two factorial ANOVA followed by paired Student’s t-test to
compare SCBF and PS values. Differences were accepted to be statistically significant at p <
0.05. All errors are given as standard error of the mean.
Figure 4.2 (See next page): (A) Representative arterial region of interest at T3 where a
cross section of the entire rat is shown. The box outlines the inset on right showing
anatomical landmarks, including the contrast filled catheter inside the jugular vein, the
aorta (used for determining the arterial input function), the trachea, and the intact T3
vertebra. The corresponding arterial input function is shown below the inset. (B) A
representative region of interest near the injury epicenter (T1/T2), and the corresponding
tissue uptake curve below. (C) A representative region of interest 0.75-1 mm rostral to
epicenter (T1) where a laminectomy was performed and the corresponding tissue uptake
curve below. (D) A representative region of interest 1.8-2 mm rostral to the epicenter
(T1/C7) and the corresponding tissue uptake curve below.
76
77
4.4 Results (Hemodynamics with FGF2)
4.4.1 SCBF and PS Measurements in the Uninjured Spinal Cord
Initial pre-injury measurements were taken at three distinct regions in the spinal cord as a
baseline comparison for all post-injury measurements. These regions were selected to
investigate the hemodynamics at the epicenter of injury and two rostral regions since the initial
mechanical impact is known to have significant biochemical and hemodynamic effects for a
considerable rostro-caudal distance. Since the lungs are directly ventral to the spinal cord in
regions caudal to the injury, respiratory movement prevented accurate measurements in these
regions. All pre-injury SCBF and PS data were averaged from all animals since treatment was
not administered until post-injury. At the region nearest to the injury epicenter (T1/T2), SCBF of
all groups pre-injury was measured at 49.7 ± 1.6 ml/min/100g. For the areas centered at T1 and
T1/C7, the pre-injury means and standard error of the means were, respectively, 45.7 ± 1.6 and
45.7 ± 1.7 ml/min/100g. These SCBF measurements for pre-injury are comparable to previously
reported values with the hydrogen clearance technique, where composite SCBF was observed to
be ~ 55-65 ml/min/100g [53].
In calculating the PS value, the assumption of the model is that Ktrans
= PS only when PS << F.
However, immediately after the trauma of spinal cord compression, there is extensive
hemorrhage into the extravascular space of the parenchyma, and therefore, this assumption does
not apply at 1 h post-injury. Thus, only pre-injury, 1 d post-injury, and 7 d post-injury PS data
are reported for each of the three regions studied. The injury epicenter at T1/T2 and the region at
T1 had identical pre-injury PS values of 0.32 ± 0.03 ml/min/g. At T1/C7, the pre-injury PS
value was calculated at 0.25 ± 0.03 ml/min/g.
4.4.2 Epicenter of Injury (T1/T2)
At 1 h post-injury and post-injection of treatment, there was a significant decrease in SCBF at
the epicenter to approximately 40 ml/min/100g for all groups (Figure 3A). At 1 d post-injury,
SCBF remained statistically lower than pre-injury values for animals that had received aCSF or
HAMC, with values of 38.0 ± 3.8 and 39.1 ± 3.5 ml/min/100g, respectively. This persistence of
decreased SCBF 1 d post-injury has been shown previously [282], and could be due to persistent
78
leaky vasculature or destruction of vasculature at the epicenter. However, animals that received
FGF2 with HAMC had a SCBF value of 47.3 ± 2.6 ml/min/100g, which was not significantly
different from pre-injury measurements. This suggests some therapeutic benefit of local release
of FGF2 in HAMC (vs. HAMC alone and aCSF alone) where blood flow has returned to near
pre-injury values at 1 d. At 7 d post-injury, all groups had elevated SCBF values compared to
pre-injury measurements; however, these were only statistically higher in those animals that
received FGF2 with HAMC and not those that received either aCSF or HAMC control
injections.
At 1 d post-injury at the epicenter, PS had increased significantly from pre-injury measurements
for all groups, and values for all groups were comparable (Figure 3B). By 7 d, PS values had
declined such that none of the groups was significantly higher than pre-injury values, but animals
that received FGF2 from HAMC had slightly lower PS values than controls. This impact on
permeability may be a reflection of improved SCBF observed in the FGF2 group at this time.
The process of angiogenesis creates leaky vasculature initially, but FGF2 is known to promote
maturation of newly formed vessels [283]. Since permeability did not decrease at 1 d for FGF2
treated animals, but SCBF did increase, it is possible that FGF2 stimulated the formation of
small vessels. At 7 d, these new vessels promoted by FGF2 at 1 d post-injury may have matured,
thus contributing to the reduced PS observed and the significant increase in SCBF.
79
Figure 4.3: (A) Spinal cord blood flow (SCBF) and (B) permeability-surface (PS) area
products nearest to injury epicenter at T1/T2 for: = FGF2/HAMC: = aCSF; =
HAMC (n= 8, mean ± standard error of the mean are shown; letters indicate statistical
differences as compared to pre-injury values where a = aCSF, b = HAMC, c =
FGF2/HAMC).
80
4.4.3 0.75-1 mm Rostral to Injury Epicenter (T1)
At 1 h post-injury and post-injection, SCBF decreased rostrally for aCSF and HAMC to 38.2 ±
3.1 and 35.1 ± 4.0 ml/min/100g, respectively (Figure 4A). Animals that received FGF2 locally
with HAMC had a SCBF of 44.1 ± 2.5 ml/min/100g at 1 h post-injury, only 1.6 ml/min/100g
lower than the pre-injury value. At 1 d post-injury, all groups showed some improvement in
SCBF, although FGF2 treated animals had values most similar to pre-injury measurements. As
was observed at the epicenter, higher SCBF was observed 7 d post-injury for the FGF2/HAMC
treated group, as compared to pre-injury measurements. A trend toward higher SCBF was also
observed for HAMC alone in this region. The aCSF group, however, only returned to pre-injury
values and did not show higher reperfusion rates at T1.
PS values rostral to the injury increased moderately at 1 d post-injury relative to pre-injury
measurements (Figure 4B), with smaller changes than those observed at the epicenter. Animals
that received FGF2 with HAMC showed a marginal increase of only 0.02 ml/min/g as compared
to the aCSF and HAMC groups, which increased by 0.11 and 0.12 ml/min/g, respectively. At 7
d, aCSF treated animals remained higher than pre-injury values, and animals that received
HAMC alone showed a reduction in PS to nearly that at pre-injury. In those animals that
received FGF2 from HAMC, the PS value at 7 d was less than that prior to injury.
81
Figure 4.4: (A) SCBF and (B) PS at 0.75-1 mm rostral to epicenter at T1 for: = FGF2/
HAMC; = aCSF; = HAMC (n=8, mean ± standard error of the mean are shown;
letters indicate statistical differences as compared to pre-injury values where a = aCSF, b =
HAMC, c = FGF2/HAMC).
82
4.4.4 1.8-2 mm Rostral to Injury Epicenter (T1/C7)
As far as 2 mm away from the injury epicenter, at T1/C7, SCBF decreased significantly from the
pre-injury measurement of 45.7 ± 1.7 ml/min/100g for all groups 1 h post-injury, to values of
38.2 ± 3.4, 36.3 ± 3.1 and 36.5 ± 1.6 ml/min/100g for aCSF, HAMC, and FGF2/HAMC,
respectively. At 1 d, all groups increased to values that were lower, but not significantly
different from pre-injury values. At 7 d, higher reperfusion rates were observed for all groups,
similar to the findings at the epicenter and T1 regions. The aCSF and HAMC treated groups
increased to 49.3 ± 2.4 and 53.5 ± 5.1 ml/min/100g, respectively. In this region at 7 d, SCBF in
animals that received FGF2 treatment was 60.9 ±3.5 ml/min/100g, which is significantly higher
than both pre-injury values and aCSF controls.
At 1 d post-injury, the PS value for the FGF2/HAMC-treated group was 0.28 ± 0.10 ml/min/g,
which was similar to the pre-injury value of 0.25 ± 0.03 ml/min/g whereas the PS values for both
aCSF and HAMC were higher, with values of 0.42 ± 0.08 and 0.37 ± 0.10 ml/min/g,
respectively. At 7 d, PS values of all groups decreased: FGF2/HAMC-treated animals showed
the greatest decrease in PS values to 0.18 ± 0.05 ml/min/g (lower than pre-injury values),
followed by HAMC-treated animals where the PS of 0.28 ± 0.07 ml/min/g was similar to pre-
injury values, and then aCSF-treated animals had PS values of 0.36 ± 0.03 ml/min/g. The
FGF2/HAMC treatment showed significantly lower PS values than aCSF-treated animals at this
time. The reduction in PS and increase in SCBF observed with FGF2 delivery further supports
the hypothesis that early delivery of FGF2 has prolonged effects on tissue recovery. The
decrease observed for delivery of HAMC alone suggests a therapeutic benefit of HAMC, which
is consistent with previous reports [10].
83
Figure 4.5: (A) Spinal cord blood flow and (B) permeability-surface area products at 1.8-2
mm rostral to epicenter at T1/C7: = FGF2/HAMC; = aCSF; = HAMC (n= 8, mean
± standard error of the mean are shown; letters indicate statistical differences as compared
to pre-injury values where a = aCSF, b = HAMC, c = FGF2/HAMC; * indicates statistical
significance between aCSF and FGF2/HAMC groups at that time).
84
4.5 Discussion (Hemodynamics with FGF2) Dynamic perfusion CT was utilized, for the first time, to serially measure SCBF and PS of the
uninjured and injured rat spinal cord. Controls showed the characteristic drop in SCBF that has
been observed with other methods [53, 282] and the spontaneous recovery that occurs within 7 d,
a period in which spontaneous angiogenesis is thought to occur [284]. PS increased following
injury as has previously been shown [52, 55], and to some extent was attenuated over time,
although it remained higher than pre-injury values at regions nearest to the site of injury.
Notwithstanding the variability in PS measurements, the results shown here suggest that dynamic
CT is an effective tool with which to study SCBF and permeability, and has the advantages of
being minimally-invasive and repeatable in the same animal.
Immediately following SCI several vascular changes occur at the site of injury. Grossly injured
vessels leak blood and fluid into the parenchyma of the tissue, leading to edema. This leads to a
disruption of the BSCB since cells and proteins normally absent in the spinal cord consequently
accumulate in the tissue. To seal burst blood vessels and limit hemorrhage, widespread
vasoconstriction occurs, thus lowering blood flow to the area. However, this global
vasoconstriction leads to ischemia in the surrounding tissue penumbra. Systemic FGF2 infusion
causes vasodilation in vivo, by opening ATP-sensitive potassium channels and enhancing release
of nitric oxide [285]. In the experiments performed here, localized delivery of FGF2 from
HAMC raised SCBF to near pre-injury values at the injury epicenter and at two rostral sites for
the first 24 h post-injury. The immediate delivery of FGF2 with HAMC in the intrathecal space
may cause vasodilation of the large peripheral vessels on the surface of the cord, potentially
leading to this stabilization of blood flow very early in the secondary injury process. In vivo
angiogenesis requires persistent vasodilation, which allows existing vessels to expand into poorly
vascularized tissue. Previously we showed that FGF2 delivered locally from HAMC was
observed to diffuse into the spinal cord for up to 6 h [239]; this localized delivery may have
contributed to the higher perfusion rates observed at 7 d post-injury at all three tissue sites, where
FGF2 concentrations sustained over several hours within the cord could have persistently dilated
small vessels, thereby stimulating angiogenic pathways and increasing blood flow relative to
controls. Moreover, FGF2 may directly stimulate angiogenesis by inducing endothelial
proliferation, or stimulating expression of VEGF [194] or nitric oxide [286] in endothelial cells
and astrocytes, respectively, to regulate angiogenesis. While the mechanism is not well-
85
understood, it is clear that notwithstanding the brief period of FGF2 delivery following SCI, long
lasting benefits were observed for SCBF after its localized delivery from HAMC.
The control injections of HAMC and aCSF showed a characteristic decrease in SCBF within the
first hour, and remained lower than pre-injury values at 24 h. Blood flow at 7 d for these
controls showed SCBF returned to pre-injury measurements, although HAMC showed a non-
statistical trend toward greater reperfusion. HA is known to scavenge free radicals [287], which
cause tissue damage and are present after SCI. HA has also been shown to promote angiogenesis
in wound healing processes by upregulating collagen production in endothelial cells [288, 289].
HA has been observed to diffuse from out of the HAMC gel within 24 h after injection [240],
although a significant improvement in locomotor function was observed at 7 d post-injury for
rats injected with HAMC relative to aCSF controls [10]. The limited benefit of HAMC alone
observed for SCBF rostral to the epicenter of injury may have contributed to the improved
functional behavior previously reported. Taken together, these results suggest that HAMC itself
may have an angiogenic impact on the tissue, which can be enhanced by the addition of FGF2.
Disruption of the BSCB occurs rapidly at the site of injury due to mechanical insult and the
downstream biochemical pathways lead to similar disturbances in the areas surrounding the
injury. The tight junctions between endothelial cells that line blood vessels and normally
effectively block passive transport of large molecules across the BSCB into the parenchyma
[290], are disrupted by injury allowing indiscriminate extravasation of plasma proteins [52]. The
BSCB remains permeable to very small molecules for at least 28 d after SCI [55], although for
larger molecules, the barrier is restored at 14 days [52]. Increased permeability of the BSCB has
been observed several vertebral segments away from the epicenter of injury [52, 55], suggesting
a widespread biochemical mechanism. FGF2 has been shown to play a role in the BSCB, where
tight junction proteins are preserved in organotypic cortical slices when exposed to FGF2 in the
surrounding media [190]. Moreover, the tight junction proteins occludin and zona occludens-1
were reduced in FGF2 knockout mice [191], suggesting a role for FGF2 in the regulation of
these proteins. FGF2 signaling has also been shown to mediate signaling between astrocytes and
endothelial cells in angiogenesis [16], and these interactions are also necessary for functioning
of the BSCB [23, 291]. HAMC alone showed a trend toward reducing PS values at regions
rostral to the epicenter of injury, and localized delivery of FGF2 with HAMC restored PS near to
pre-injury values in these same regions 1 d post-injury. At 7 d post-injury, the PS values were
86
lower than pre-injury values and significantly lower in animals treated with FGF2/HAMC than
those treated with aCSF in the region most rostral to the epicenter at T1/C7. These data also
support the hypothesis that short-term delivery of FGF2 with HAMC can have a long-term
impact on tissue recovery.
Here we have shown that SCBF measurements at three distinct regions of the spinal cord show
characteristic patterns of recovery using dynamic perfusion CT imaging. Localized delivery of
FGF2 from HAMC stabilized SCBF nearest to pre-injury values, also causing increased
perfusion rates at 7 d post-injury as compared to pre-injury measurements. PS measurements
indicate a characteristic increase in permeability 1 d following injury. Delivery of FGF2 with
HAMC stabilized PS at regions rostral to the epicenter of injury, with significantly lower PS
values than aCSF controls at 7 d in the region most rostral to the epicenter at T1/C7. HAMC
alone also showed a moderate improvement in PS at regions distal to the epicenter, supporting
previous data that the delivery vehicle itself may have therapeutic benefit. Importantly, we
report here that dynamic perfusion CT is a valid method for serially measuring SCBF and
permeability after SCI in small animals, and that FGF2 delivered locally with HAMC can
improve SCBF and permeability following injury.
87
5 Discussion
5.1 Localized Drug Delivery Spinal cord drug delivery is particularly difficult due to the layers of protection surrounding the
CNS, which effectively limits molecular diffusion into healthy tissue. After SCI the BSCB is
compromised due to grossly injured vessels from the mechanical impact, which allows blood and
any systemically administered drugs passage into the parenchyma. Intravenous delivery relies
on this BSCB disruption to achieve a therapeutic impact; however, the high doses required to
accomplish this can cause significant systemic side effects. Thus, localized delivery strategies
have been developed to improve therapeutic concentrations while minimizing systemic impacts.
Intrathecal injections can be done, but provide only transient delivery at the site of injury, and
may be followed by leakage of CSF after injection. Minipumps can constantly deliver drugs in
the epidural or intrathecal spaces for long periods of time when implanted locally, allowing high
local concentrations near the site of injury. Although minipumps and catheters provide direct
access, they also can cause scarring and compression of the spinal cord, which is a barrier for use
post-SCI where further damage of the cord may only worsen the potential for functional
recovery. Polymeric delivery systems have also been used for localized delivery. Fibrin glues
have also been applied extradurally to deliver drugs [148]; however, since the dura/arachnoid
mater generally only allow passage of small molecules (<400 Da) [200], proteins or larger
molecules require transport moieties to gain access to the tissue. Additionally, molecules must
also traverse the intrathecal space where CSF flow can disperse the drug before reaching the
target tissue area. Combining the concepts of a minipump, which could sustain drug release, and
an intrathecal bolus injection, where a small needle could be inserted and removed immediately
after delivery, an injectable collagen matrix was developed for localized and sustained
intrathecal delivery [9, 87]. A blend of 2 wt% HA and 7 wt% MC was subsequently developed,
which was non-cell adhesive, unlike collagen, and also appeared to seal the dura from CSF
leakage [10]. These systems were shown to be safe for intrathecal drug delivery, without the
scarring and compression that resulted from catheters and minipumps. The following discussion
focuses on work done to better understand the capacity of HAMC as an injectable intrathecal
drug delivery system.
88
5.1.1 In vitro Release of Therapeutic Molecules from HAMC
Numerous molecules have been shown to provide therapeutic benefit in animal models of SCI,
thus in vitro release of several molecules from HAMC was investigated. EPO is a 30 kDa and
highly glycosylated protein which has shown promise for improving functional recovery after
SCI. FGF2 is an 18 kDa protein which can have a spectrum of actions after SCI, including
neuroprotection and angiogenesis. In addition to these proteins, two small molecules were also
investigated: 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzoquinoxaline-2,3-dione (NBQX) and
nimodipine. NBQX is a glutamate antagonist that may provide neuroprotection from glutamate
excitotoxicity following SCI [118]. It is a small molecule of 336 Da which is quite hydrophobic,
but the sodium salt form is readily soluble in aqueous medium. Nimodpine is a calcium channel
blocker of 418 Da, generally used for treating vasospasm, and has been tested clinically for SCI
[110]. It is highly hydrophobic and is only sparingly soluble in aqueous solutions, where the
reported solubility is 2 – 4 µg/ml [292]. The release of these 4 molecules from HAMC was
studied to better understand the potential of HAMC as a drug delivery system for neuroprotective
molecules to be delivered early in the secondary injury process.
Both EPO and FGF2 showed similar release profiles from HAMC, although FGF2 is not
glycosylated. Nearly 99% release of both molecules from HAMC occurred within 16 h in vitro,
when the volume of HAMC to aCSF was similar to in vivo ratios. To understand how molecules
would be released from HAMC in vivo, the mechanism for drug release was investigated in vitro
where some variables such as polymer geometry could be controlled. Models that define the rate
of drug release from drug delivery systems are based on Fick’s laws of diffusion, where the rate
of change of the concentration, C, over time, t, is related to the rate of concentration change over
its position, x, within the drug delivery system at that time, given by:
where at and at ; D is the diffusion
coefficient of a given molecule. This assumes perfect sink conditions whereby drug diffuses
from a thin polymer slab at an initial uniform concentration, C1, and the surrounding media into
89
which the drug is diffusing has zero concentration, or C0. While numerical solutions can be
obtained for this complex partial differential, simplified empirical models have been established
by Peppas [253] where rate of drug release, Mt/M∞, is related to time and distance through:
where l is the thickness of the polymer that the molecule must diffuse through. This is valid for
the first 60-80% of Fickian release (Mt/M∞ ≤ 0.60 – 0.80), and assumes a simple slab geometry
from which drug can be released from planar surfaces of the slab. This can be generalized to:
Where k is a constant that takes account of the 4(D/πl2) 1/2
Figure 2.1
term in the previous equation and n is
the diffusional exponent indicative of the transport mechanism. Still utilizing the same
assumptions, an n = 0.5 indicates Fickian diffusion and n > 0.5 indicates non-Fickian diffusion,
which is the case in swelling or degradation controlled release systems. However, it is important
to note that deviations from one-dimensional release and simple polymer geometries without
sink conditions could impact the interpretation of both k and n, thus the method by which the
release mechansism is studied should be as close as possible to a polymer slab with a surface
exposed to media not containing drug for release. In vitro release of both EPO and FGF2 were
obtained with similar geometries and controlled sink conditions by refreshing the media in which
the HAMC containing drug was maintained. B shows that cumulative EPO release
increases linearly with t0.5 for up to 75% of release, indicating that its release is diffusion
controlled. Similarly, release of FGF2 and PEG-FGF2 from HAMC showed the same pattern of
linearity with t0.5 Figure 0.1(Appendix 3, ), suggesting that protein release from the physically
interacting gel network of HAMC is diffusion controlled. This would likely hold true for
proteins of approximately the same molar mass that do not interact with the hydrogel polymers.
Doing these experiments in a controlled, in vitro experiment allowed us to understand better how
HAMC would deliver drugs in the more complex in vivo environment, where geometry is not
simple, and ideal conditions cannot necessarily be maintained.
90
The release of any molecule from a drug delivery system depends on both the delivery vehicle
and the diffusion rate of the drug molecule, which is encompassed in the k term above.
Diffusivity depends upon several factors, including size, charge, and solubility. The different
slopes, indicative of the k term above, observed between EPO and FGF2 demonstrate that they
have different diffusivities in HAMC. Diffusive release was also observed for the small, water
soluble NBQX, where complete release was observed within 1 h in vitro and >85% of
cumulative release was linearly related to t0.5 Figure 0.2 (see Appendix 3, ). Additionally, a
soluble formulation of nimodipine in HAMC containing 2% v/v ethanol followed a similar trend
to NBQX and protein release, where complete release occurred within 8 h in a diffusion-
mediated manner (Appendix 3, Figure 0.2). However, for the sparingly water-soluble
nimodipine, both dissolution of the drug and its diffusion through the matrix determined the rate
of drug release [293, 294]. Solid particles of nimodipine were produced with varying degrees of
sonication to achieve crystalline particles of approximately 900 nm and 380 nm, and these were
incorporated into HAMC. The solid particles of nimodipine followed a biphasic release pattern
with a high burst release within hours and a slow, sustained release for several days (Appendix 3,
Figure 0.3). The high initial burst release in 4 h for 380 nm particles was 80% and 60% for the
900 nm particles. The 380 nm particles had a secondary release phase for 2 d, and the 900nm
particles for 3d. This biphasic release suggested that HAMC somehow impacted the solubility of
nimodipine. The presence of hydrophobic moieties on MC was thought to increase the intrinsic
solubility of solid drug particles in aqueous systems, since polymeric excipients similar to MC,
such as hydroxypropyl methylcellulose or poly(vinylpyrrolidone), have been incorporated into
drug particles to increase their solubility [295-299]. This was verified by mixing nimodipine
particles in HA and MC separately, revealing that MC, but not HA, increased the solubility by an
order of magnitude from 2-4 µg/ml to 30-40 µg/ml (Appendix 3, Figure 0.4). Ultimately, these
studies showed that HAMC accelerated the release of a hydrophobic, poorly soluble drug, also
yielding highly tunable release profiles dependent on the size of solid particles. These studies
with nimodipine are described more fully in [300], where numerical predictions of release were
matched to this empirical data showing that drug release was dissolution controlled rather than
diffusion controlled.
In vitro data shows that proteins and water soluble small molecules are released from HAMC in
a diffusion-controlled manner. Release of nimodipine, a sparingly soluble drug showed that
91
longer delivery could be achieved through a drug dissolution controlled mechanism. The
diffusive mechanism of release for proteins suggested that in vivo delivery of these molecules
with HAMC would likely be limited to 12-16 h post-injection. Hence, neuroprotection rather
than neuroregeneration was targeted for events that occured in the immediate secondary injury
phase, specifically, the ischemic insult that results from vascular damage, hemorrhage, and
edema. Notwithstanding this short delivery period, intrathecal HAMC delivery has potential for
longer lasting effects since the apparent protein concentrations on the surface of tissue can
remain more than 100 times longer after injection when delivered with HAMC than conventional
methods such as intrathecal bolus or intraperitoneal delivery, based on the average velocity of
CSF flow [12].
5.1.2 In vivo Delivery from HAMC
5.1.2.1 Delivery Period and Degradation of HAMC
In vitro data showed that the HAMC gel degraded within 15 d [10] but little was known about
how long the gel lasted in vivo where degradation could be affected by changes in pH, gel
geometry, and the presence of enzymes. Each component of HAMC was thus fluorescently
labeled to observe the presence of it in the intrathecal space at the site of injection over time.
Figure 2.2C shows the in vivo degradation of each component, showing that 95% of HA
disappears from the gel within 24 h, and nearly 60% of the MC component is lost within that
time as well. The structure of HA contains hydroxyl, carboxyl and amino groups, which are
hydrophilic. MC contains some hydrophilic hydroxyls as well as the hydrophobic methyl
groups. At low temperatures these polymers are simply entangled together, but after injection
into the intrathecal space, the temperature increase leads to associations between MC molecules
when the cagelike water structures around the hydrophobic methyl groups are broken. The blend
of HAMC thus becomes aphysically crosslinked network, in which MC molecules form a
network in which the HA is entangled. The blend of HAMC in its crosslinked state is thus a
semi-interpenetrating network, whereby MC cannot dissolve out without breaking bonds, but the
disentanglement of HA from the gel does not require any bond disruption. The snake-like
motion of HA out of MC is termed repatation, in which the macromolecule changes
configuration as it withdraws from the gel phase. HA may be removed through this process, or
92
may undergo enzymatic degradation by hyaluronidases secreted locally following an injury [301]
or entering the cord from hemorrhaging blood. In addition, HA can also be non-specifically
degraded by reactive oxygen species, particularly after injury [302, 303]. This early loss of HA
from the gel structure likely leads to a more porous network of MC, where CSF replaces the void
space left by the dissolving HA. By making the gel more porous, diffusive drug release may be
expedited since the diffusion path of drug is less tortuous through larger pores. While early loss
of HA may lead to faster drug release, nearly 30% of MC continued to persist in the intrathecal
space for up to 4 d after injection, with complete degradation by 7 d. These results indicate that
HAMC permits a maximum release window of 4-7 d in vivo, if drug release could be extended
beyond diffusion mediated release. Importantly, these results indicated that the gel was still
present in the intrathecal space during the time window of in vitro release, suggesting that
localized delivery in vivo could be achieved.
In vitro release of EPO from HAMC was found to be diffusion controlled, which suggested that
HAMC was capable of releasing EPO for up to 16 h. These experiments were supported by the
estimation of 10 h based on gel thickness and diffusion coefficient of EPO. Thus, the residence
time of EPO on the surface of the spinal cord in contact with the gel could be estimated at 10 h.
An intrathecal bolus of EPO, however, was expected to be washed away more quickly from the
site of injection by the flowing CSF. The residence time at the site of injection for EPO
delivered IT was estimated to be on the order of 5 min. This estimation was obtained using the
same volume of injection as for EPO delivered from HAMC (10µl), but the CSF flow rate was
used rather than the diffusion coefficient. This was then used to show the ratio of residence
times for EPO delivered from HAMC and as a bolus IT: . This ratio shows
that EPO delivered with HAMC could result in more than a 100 fold increase in the residence
time of EPO delivered as an IT bolus, when only the site of injection is considered. These
estimations suggested that EPO delivered with HAMC would provide better diffusion into the
injured tissue at the site of injection.
In vitro release of both FGF2 and PEG-FGF2 from HAMC was diffusion controlled, so
experiments were performed to validate that proteins diffusing out of HAMC actually penetrated
into the injured spinal cord rather than simply diffusing into the CSF. This was done by
93
sectioning and digesting spinal cords serially after injecting HAMC/FGF or HAMC/PEG-FGF2
into the intrathecal space. Using ELISA to detect the protein in the tissue digests, FGF2 and
PEG-FGF2 were confirmed to diffuse into the spinal cord (Figure 3.6, B and C). Concentrations
of 5-50 ng/ml were achieved within the tissue, where the highest concentrations were observed
in the outer 600 µm of tissue, nearest to HAMC. These are concentrations that are similar to
culture experiments where 25-50 ng/ml of FGF2 in ex vivo brain slices showed reduction in
permeability of the blood-brain barrier [190], 5-40 ng/ml is used in differentiation/proliferation
media for stem cells [304], and neuronal cultures could be spared from glutamatergic
excitotoxicity with 20 ng/ml [188]. These results verified that relevant tissue concentrations
could be achieved using HAMC as an intrathecal delivery system.
While relevant concentrations were achieved within the tissue, the cumulative amount of protein
that was delivered to the tissue relative to the amount loaded into HAMC was quite low, where
only 0.5-1% of FGF2 loaded into HAMC penetrated into the tissue. One possibility is that much
of the FGF2 diffused from HAMC into the CSF very quickly. This could account for the higher
concentrations observed in the center of the cord, where FGF2 released from HAMC into the
CSF may have circulated through the central canal of the spinal cord. However, based on a
diffusion mechanism it is unlikely that 99 % of the FGF2 diffused from the gel so quickly, since
the CSF-polymer interface is small compared to the spinal cord-polymer interface. Additionally,
if so much of the FGF2 had recirculated through the central canal, higher concentrations would
have been expected in the center of the cord. FGF2 is highly regulated in vivo by its binding to
heparin in the extracellular matrix, where the FGF2-heparin Kd = 39 nM [305]. The rate of
association for these molecules is quite fast, whereas the rate of dissociation is quite slow.
Although soluble heparin was delivered with FGF2 in HAMC to limit its sequestration in the
extracellular matrix, the FGF2-heparin complex can also bind to FGF Receptors 1-4 (FGFR1-4),
with FGFR1 being the most highly expressed receptor. The FGF2-FGFR1 complex has a Kd =
62 nM [305], and similar to FGF2-heparin complexation, this has a fast rate of association and a
slow dissociation rate [306]. FGF2 detected using an ELISA would only detect FGF2 that was
not bound to receptors on the tissue, thus, the low concentrations observed in the spinal cord
would only reflect diffusing FGF2. This suggests that tissue concentrations may have been
higher if FGF2 bound to receptors or to heparin in the extracellular matrix could have been
detected with the ELISA.
94
While FGF2 and PEG-FGF2 penetrated into the spinal cord at relevant concentrations, it was
only detected in the spinal cord for up to 6 h, and was not observed in the tissue 24 h after
HAMC injection. The exponential decrease in concentration from the polymer surface observed
at 3h (Figure 3.6B) was beginning to plateau at 6 h (Figure 3.6C), suggesting drug depletion in
HAMC. Since an intrathecal bolus injection would be transient at the tissue surface due to CSF
circulation, HAMC effectively sustains the delivery of proteins to the injured spinal cord for at
least several hours. Although this time period is short with respect to the secondary injury, it
may be an effective time period to limit tissue damage that occurs in this timeframe. This early
delivery is sufficient to have some therapeutic benefit, which is further discussed in sections
5.2.1 and 5.2.2 of this chapter, where neuroprotection was observed with EPO delivery, and
increased blood flow in tissue was observed for FGF2 delivery even at times when diffusing
FGF2 was no longer detected. However, as evidenced from EPO administration, the functional
benefit that results from early delivery may be limited, thus a longer delivery window may be
beneficial and necessary to provide greater neuroprotection after some of the early secondary
injury events have subsided.
5.1.2.2 Potential Biological Role of HA In vivo
HAMC injected into the intrathecal space interfaces with the dura/arachnoid maters, spinal cord
tissue, and CSF. Polymer dissolution was expected to occur mainly into the CSF since the long
chains would face the lowest diffusive barrier in a fluid. The spatial limitations imposed by the
dura/arachnoid and the spinal cord allowed only for rostro-caudal polymer dissolution. In
assessing the safety of HAMC, significantly better functional recovery was observed in animals
that received HAMC over the aCSF injection [10], several days after HA had been cleared from
the intrathecal space. Also, the hemodynamic characeteristics of HAMC treated groups showed
trends toward improved SCBF in sites rostral to the epicenter [241]. These results suggest that
HAMC plays a biological role following injury.
HA comprises a significant amount of the extracellular matrix in the CNS, and in healthy tissue
is of a high molecular weight form (>3-6 x106 Da). However, after an injury, HA is believed to
play a central role in the wound healing process, whereby the high molecular weight form that is
normally found in the extracellular space is broken down locally by secreted enzymes. These
95
lower molecular weight products then have multiple actions through each phase of the wound
healing process (reviewed in [307]). Initially, HA enhances immune cell infiltration and
upregulates inflammatory cytokines [308, 309]. In addition to these early actions on cells, HA
scavenges free radicals and has antioxidant properties [287, 310]. As the high molecular weight
form in the tissue is cleaved, HA fragments of <2x105
5.1.2.3 HAMC Delivery Compared to Conventional Delivery Strategies
Da have been observed to induce
inflammatory or proliferative genes [311], which is thought to contribute to cell migration for
tissue remodeling. These lower molecular weight forms are also thought to enhance the
angiogenic processes in the later phases by upregulating collagen production in endothelial cells
[221, 288, 289]. While it is unlikely that the high molecular weight, exogenously administered
HA in the intrathecal space diffuses rapidly into spinal cord tissue to increase the cellular
migration into the injured zone following SCI, it is possible that HA scavenges free radicals
which diffuse into the intrathecal space from the injured tissue, and that secreted enzymes also
lead to its degradation. The lower molecular weight fragments of HA that result from oxidation
or enzymatic degradation may diffuse into the tissue or recirculate in the CSF through the central
canal. Potentially, this increase in low molecular weight HA could lead to enhanced
angiogenesis as the acute inflammatory response begins to subside, thus accounting for the
enhanced blood flow response [241], greater neuron counts [240], and motor function [10]
observed in the HAMC alone groups.
HAMC was compared to intraperitoneal delivery and intrathecal bolus delivery to investigate the
potential benefits that could be achieved with localized and sustained delivery. Intrathecal
HAMC delivery was calculated to sustain release at the site of injury more than 100 times
intrathecal bolus or intraperitoneal delivery, and systemic delivery via intraperitoneal injection
was expected to acheive the lowest dose of EPO at the site of injury since it was diluted through
the entire bloodstream. As expected, this route of delivery did result in the lowest amount of
neuroprotection relative to intrathecal HAMC delivery of EPO (Figure 2.4). Similarly, the
intrathecal bolus resulted in fewer surviving motor neurons relative to intrathecal HAMC
delivery of EPO, likely due to the fast clearance by CSF flow. This improvement in neuron
counts may be attributed to longer residence time and bioavailability of EPO released from
HAMC at the site of injury. Thus, these studies established that EPO delivered locally with
HAMC increased neuronal sparing as compared to more traditional routes of drug delivery.
96
In addition to providing enhanced neuroprotection, HAMC delivery may have also provided
other benefits over the other two delivery strategies. Because EPO causes red blood cell
production [261] thus leading to blood thickening, an intravenous route of administration might
lead to more severe hemorrhaging after injury. However, a more significant impact was seen
between the two localized delivery strategies of HAMC containing EPO versus intrathecal bolus
of EPO. In animals that received the bolus intrathecal delivery of EPO, acute respiratory
complications occurred during surgery, where deep gasping sometimes led to death. Of the 16
rats that underwent surgery in this group, 12 exhibited this behavior, 4 of which could not be
resuscitated and died. Slowed breathing has been reported for EPO when it is administered in
the central nervous system [262], yet these studies indicate more serious physiological
implications for intrathecal bolus delivery. While this impact was not observed in animals that
received intraperitoneal EPO or intrathecal HAMC alone, deep breathing was observed in 2 of
10 animals that received intrathecal HAMC delivery of EPO, both of which were successfully
resuscitated. This unexpected respiratory complication may have contributed to the lack of early
functional recovery after the first week in those animals that received EPO as an intrathecal
bolus, where interestingly, motor function was statistically lower than those that received EPO
from HAMC intrathecally. These observations demonstrate that HAMC may have an added
benefit in terms of improved safety for intrathecal drug administration.
5.2 Neuroprotection Achieved with Localized Delivery The secondary injury cascade, which leads to significant tissue damage and ultimately the
neurological deficit after SCI, begins quite rapidly after the initial mechanical trauma. Limiting
the early events that eventually lead to widespread cell death and neuronal desctruction has been
investigated with a variety of pharmacological treatments. Neuroprotection is a term describing
neuronal sparing and reduced lesion volume following SCI, and attempts to achieve this largely
target early secondary secondary injury events. Localized delivery with HAMC permits release
of neuroprotective proteins at the site of injury early in this process; here, the tissue impacts of
early delivery of EPO and FGF2 with HAMC are described.
97
5.2.1 Implications of EPO Crossing the BSCB
EPO delivered from HAMC provided enhanced neuronal protection, likely due to direct contact
with EPO receptors on the surface of neurons [155, 159]. Cultured primary neurons treated with
EPO have been protected from excitotoxic cell death [155], and experiments with EPO in stroke
and traumatic brain injury models [170] have also shown a neuroprotective effect in hypoxic and
ischemic environments [169]. More indirect actions of EPO suggest that it can reduce the
inflammatory response following traumatic brain injury [163, 263, 264] and can generally
decrease apoptosis [163, 264]. When delivered from HAMC, EPO showed greater functional
benefit at 7 d post-injury over EPO delivered IT, although this impact was not retained at later
times. Notably, none of the delivery methods of EPO achieved the same high level of functional
recovery that was previously reported [49]; however, an attempt at replicating this work in a
different lab showed that no significant functional recovery could be achieved with EPO
administration [312], thus validating the lack of sustained fuctional recovery that we observed
even with the same route of delivery as the original work.
The spatial limitations of HAMC in the intrathecal space results in a substantial hydrogel-spinal
cord interface, whereby drugs are delivered in a sustained manner for several hours to the spinal
cord surface. The drug delivered to the surface of the cord must diffuse through the pia mater
and through spinal cord tissue to reach cell receptors. For an intrathecal bolus delivery method,
this would also be the case; however, since it disperses into the CSF, drug concentrations would
be significantly lower than those at the HAMC-spinal cord interface. Since EPO crosses the
blood-brain barrier [172], it may in fact have a shorter diffusion distance when administered
systemically, since it must only diffuse from the luminal surface of the endothelium into tissue.
For highly vascularized tissue, particularly in the spinal cord where a dense capillary network
exists [313], a significant surface area of endothelium could potentially allow EPO to reach the
ventral motor neurons faster than EPO delivered via HAMC. This is substantiated by the fact
that neuron counts were nearly equivalent for EPO delivered IP or IT. However, local delivery
of EPO with HAMC still resulted in greater motor neurons counted than both of these groups,
suggesting that the local dose when EPO was sustained at the site of injury for several hours was
greater than the dose that could be delivered with transient delivery intrathecally or systemically.
98
5.2.2 Improvements in SCBF and Permeability with FGF2 Delivery
FGF2 is a molecule that does not cross the BSCB [314] and since FGF2 is mitogenic [182, 197],
it was not delivered systemically, thus localized delivery of FGF2 from HAMC was compared to
controls of HAMC alone and aCSF intrathecally. Immediately following an injury, several
vascular changes occur at the site of injury. Vessels that are grossly injured will leak blood and
fluid into the parenchyma of the tissue, leading to hemorrhage and edema. This leads to a
disruption of the BSCB since cells and proteins normally absent in the CNS consequently
accumulate in the tissue. Immune cells, such as leukocytes, infiltrate the tissue via hemorrhage
and through the leaky vasculature, stimulating a significant immue response and emitting
reactive oxygen species and other free radicals. To seal the burst blood vessels and limit the
hemorrhage, vasoconstriction occurs, thus lowering blood flow to the area. However, this global
vasoconstriction leads to ischemia in the surrounding tissue penumbra. Following SCI, localized
delivery of FGF2 from HAMC stabilized SCBF near pre-injury values at the injury epicenter and
at 2 rostral sites for the first 24 h post-injury. Also, at 7 d post-injury perfusion rates were
increased beyond pre-injury measurements at all 3 sites observed. The control injections of
HAMC and aCSF showed a characteristic decrease in SCBF within the first hour, and remained
lower than pre-injury values for up to 24h. At 7 d, SCBF was statistically higher than pre-injury
values for FGF2 treated animals. Blood flow was also higher for controls, but not statistically
higher than pre-injury values, suggesting that FGF2 had a greater impact on the tissue. In vivo
angiogenesis requires persistent vasodilation; however, this impact has been difficult to link to a
biochemical process [286]. Systemic FGF2 infusion has been shown to cause vasodilation in
vivo by opening ATP sensitive K+ channels and enhancing release of nitric oxide [285].
Although FGF2 was only detected in the spinal cord for up to 6 h in vivo, it may have caused
early vasodilation on the large vessels that are on the periphery of the cord, since it would have
reached these vessels first, based on in vivo diffusion data for FGF2. In areas distal to the
epicenter where localized FGF2 delivery with HAMC appeared to stabilize blood flow at 1 d
post-injury, and higher perfusion rates at 7 d post-injury, several possibilities exist. One
possibility is that, as described above in section 5.1.2.1, exogenously delivered FGF2 was
potentially sequestered in the extracellular matrix or bound to cells, and could continue to have
actions much later than the expected delivery window. Another possibility is that FGF2
diffusing into the tissue several hours after the injury would cause earlier vasodilation of smaller
99
vessels than controls, at a time when hemorrhage was beginning to be controlled. This early
vasodilation of smaller vessels could have stimulated angiogenic pathways earlier than controls,
thus increasing blood flow at later time points. Additionally, FGF2 may have directly stimulated
angiogenic impacts in cells at these early times, either by inducing expression of VEGF [194] or
nitric oxide [286] in endothelial cells, which are known angiogenic regulators. Based on cerebral
blood flow studies, capillary density is directly correlated to cerebral blood flow [315]. If
capillary density was increased due to angiogenesis, then increased blood flow would be
expected. Although the extent of angiogenesis required to produce a beneficial impact on tissue
sparing and functional recovery is not known, the capillary density of the uninjured spinal cord
could be used as a baseline goal for a given vertebral segment. While the mechanism for
improved SCBF observed here was not fully elucidated, FGF2 did provide long lasting benefits
on SCBF after its localized delivery.
Disruption of the BSCB occurs quite quickly from the mechanical impact at the site of injury,
although the molecular mechanisms by which disruption occurs in the rostro-caudal regions of
the injury are not completely understood. The inner lumen of blood vessels in the CNS has a
glycoprotein-rich, anionically charged glycocalyx coating, providing a repulsive barrier to
plasma proteins of similar charge. The tight junctions between endothelial cells that line the
vessels effectively block passive transport of large molecules across the BSCB [290], and these
are surrounded by a basement membrane abluminally. Astrocytes surround the basement
membrane to further regulate what is passed into the parenchyma. Following injury, the
glycocalyx transiently loses anionic charges leading to indiscriminate extravasation of plasma
proteins [52], likely also regulated by tight junction disruptions. After SCI, the BSCB remains
permeable to very small molecules for at least 28 d [55], although larger molecules are
permeable for closer to 14 days [52]. Permeability of the BSCB has been observed several
segments away from the epicenter of injury [52, 55], suggesting a widespread biochemical
mechanism. Localized delivery of FGF2 with HAMC stabilized PS at regions rostral to the
epicenter of injury at 1 d post-injury. At 7 d post-injury, FGF2 lowered PS values with respect to
pre-injury measurements, showing significantly lower PS values than aCSF controls at 7d in the
region most rostral to the epicenter at T1/C7. Knockout mice for FGF2 showed reduced levels
of tight junction proteins in the CNS [191], suggesting that FGF2 regulates their expression. In
organotypic brain slices, significantly higher expression of tight junction proteins and intact
100
blood vessels persisted for 7 d in vitro when bathed in media containing FGF2 [190].
Additionally, the interaction of astrocytes and endothelial cells are necessary for functioning of
the BSCB [23, 291]. In angiogenesis, paracrine signalling leads to changes in both astrocytes
and endothelial cells [16], thus localized FGF2 may have impacted permeability due to its effects
on either astrocytes or endothelial cells.
5.3 PEGylation for Improving Tissue Delivery Since prior experiments showed that FGF2 did not significantly penetrate injured spinal cord
tissue [87], PEG was conjugated to FGF2 with the goal of enhancing tissue delivery.
PEGylation has previously been used to improve molecular stability [234], allowing for greater
bioavailability. It also reduces immunogenicity [236] by shielding immunogenic sites of
proteins, which can lead to decrease tissue clearance [237]. PEG has increased the biologic half-
life of several molecules and also enhanced distribution in tissues [238]. The distribution of
PEGylated proteins in tissue depends on both the physico-chemical factors of the PEG-protein
conjugate and the physiological and anatomical factors of the host tissue. While PEG increases
the size and molecular weight of a protein, it also can produce changes in steric hindrance, the
electrostatic binding properties, hydrophobicity, and the PI of the protein. All of these changes
may impact its bioactivity and action once in vivo. Additionally, the physiological factors of the
tissue in which a PEGylated protein is diffusing will impact the effectiveness of PEGylation.
FGF2 is a polypeptide that retains its conformation mainly through van der Waals forces and
hydrogen bonding [175]. Although it contains 4 cysteine residues, none of them participate in
disulfide bonding [178]. Heparin and heparin sulfates bind to FGF2 and make them
conformationally more stable and able to bind to FGFRs. This suggests that FGF2 alone does
not have an extremely stable tertiary structure and may be easily rendered inactive, thus both
FGF2 and PEG-FGF2 were co-delivered with heparin. When PEG was attached to FGF2, in
vitro bioactivity did not decrease, suggesting that it retains its ability to attach to cell receptors.
Two interesting findings came from the localized delivery of FGF2 intrathecally with HAMC: 1)
FGF2 does in fact penetrate into spinal cord tissue, but previous methods were not sufficiently
sensitive to detect it; and 2) PEGylation of FGF2 improves its diffusion into tissue. Assuming
that PEG-FGF2 retains the ability to bind to receptors, we showed that PEGylation increased the
101
concentration of FGF2 in the injured spinal cord tissue by nearly twofold. Diffusivity of PEG-
FGF2 was expected to decrease relative to FGF2 due to the increase in molecular size, which
would cause it to take a more tortuous path through the tissue. However, PEGylation was
thought to also decrease the elimination of FGF2 diffusing through the tissue. While each
parameter could not separately be determined from the concentration profiles, the inverse
relationship between diffusivity and elimination, given by , could
empirically be determined based on these data. The ratio was expected to be
<1 if the φ value reflected only a change in diffusivity, assuming elimination remained constant.
However, instead, the ratio , suggesting that PEGylation of FGF2
caused significantly reduced elimination of the protein. Thus, the greater tissue penetration
observed for PEG-FGF2 was ascribed to reduced elimination.
5.4 Spinal Cord Blood Flow Imaging Strategies The primary goal of this thesis was to achieve localized neuroprotective drug delivery to the
injured spinal cord. However, since FGF2 mechanstically impacts vasculature in tissue, a
method to functionally assess blood flow and permeability was desired to better understand its
local delivery. Several methods have been utilized for imaging SCBF, including xenon [60, 61]
and hydrogen [53, 62] clearance methods and autoradiographical methods [50, 64]. All of these
methods are limited to either single session or terminal measurements of SCBF, leading to a
large quantity of animals needed. Also, since variability between animals can be high for such
measurements, several measurements of SCBF in the same animal were desirable within 1 w,
which is the time frame when post-traumatic angiogenesis occurs. We first investigated
ultrasound imaging paired with laser Doppler to image blood flow through spinal cord tissue.
This method seemed promising, since an external probe applied to the muscle tissue over the
cord could be used for measurement. Measurements from individual vessels of artieral and
venous flow could be observed in the healthy spinal cord using this method (Appendix 3, Figure
0.5), but several difficulties with this method were encountered. Although the probe could be
102
placed on the skin surface, high frequencies were necessary to probe as far as the spinal cord,
and much of the signal was lost due to the shadow of the laminae that were not removed. Other
groups that do this type of imaging apply the probe directly to the spinal cord surface, with the
muscle and fascia removed [278, 316], which is invasive and impractical. For this reason, we
pursued CT imaging to study blood flow dynamics, a technique which also allowed us to study
permeability.
CT imaging does not require a probe to be applied to the tissue; instead it provides sectional
images of the whole body. When a contrast agent is injected into the bloodstream close to the
heart, the increase in density as the tracer moves through the tissue slice is detected, and the
perfusion rate can be derived from this first pass through the tissue. Based on the concentrations
that accumulate in the tissue, the permeability can be assessed, since these tracers are generally
not very permeable across the uninjured BSCB. Contrast agents used in CT are approved
clinically for blood flow assessments, and most compounds are quickly filtered from the blood
via the kidney, and therefore have low systemic toxicity. Dynamic CT imaging has been used in
several applications in both animal and human studies. Initially the method was used to
investigate renal blood flow [68, 279, 280], but further technical developments allowed
measurements of tumor blood flow in cerebral tumors [67, 73], lung cancers [72, 74], as well as
renal and hepatic metastases [68, 70, 71]. A recent report describes the first use of CT perfusion
for SCBF in humans where different software packages for calculating SCBF and permeability
were tested to investigate the reproducibility of these measurements [66]. This study showed
that CT perfusion provided reasonable reliability for spinal cord blood flow of the cervical spinal
cord region in humans. We tested this method in rats, where their small size requires greater
sensitivity and resolution in CT. Most small animal studies dealing with SCI treatements involve
histological or immunohistochemical examination of tissue, which is not practical for clinical
studies. Developing a minimally invasive tool for measuring SCBF and permeability in humans
and small animals would provide a unique ability to compare the impacts of SCI treatment
between species. Chapter 4 describes in detail the mathematical models which underlie the
calculations for SCBF and permeability measurements. The SCBF parameter is easily
determined using first-pass tissue and arterial influx parameters for contrast agents that are not
very permeable across the BSCB, such as the iodinated compound that was employed. This
method has been used for renal blood flow calculations [68, 70], and SCBF values calculated
103
using this model showed similar values in rats to those calculated using the hydrogen clearance
method [53]. SCBF decreased following injury as previously reported [50, 53, 316], and hyper-
reperfusion was observed 1 w after injury. The modified Toft’s model was identified as an
appropriate method for assessing the permeability-capillary surface area product, in which the
intravascular and extravascular spaces are viewed as separate compartments. These data showed
an increase in permeability at 1 d post-injury, which has been similarly demonstrated in other
SCI models [55, 317, 318]. These characteristic trends of blood flow and permeability over time
that have been shown using other methods support the data obtained with dynamic CT imaging,
suggesting that this is a valid measurement technique for these paramaters. Dynamic CT
imaging is an important method by which functional tissue recovery can be assessed, rather than
static measurements provided by immunohistochemical and histological data. Additionally, CT
imaging improves on existing methods for SCBF and permability measurements since it can be
done serially on the same animal, thus reducing the size of animal studies. This data and recent
advances in human SCBF measurements with dynamic CT may also provide a unique
opportunity to directly compare tissue outcomes in clinical studies with the preclinical small
animal studies that precede them.
5.5 Acheivement of Objectives
The hypothesis originally proposed for this research was:
Localized delivery of neuroprotective agents with hyaluronan and methylcellulose will
mitigate tissue damage after spinal cord injury.
This dissertation describes the use of EPO and FGF2 as neuroprotective agents to assess the
utility of HAMC for localized delivery after SCI. Below is a summary of work completed
toward achieving each of the objectives defined at the beginning of this document.
1. Characterize the in vitro and in vivo release of neuroprotective agents from HAMC
The release mechanism of proteins and water-soluble small molecules from HAMC was
determined to be diffusion mediated based on in vitro data. These results suggested that HAMC
was an ideal drug delivery system for neuroprotective agents in the acute phase of secondary
104
injury after SCI. In vivo release of FGF2 from HAMC resulted in therapeutic concentrations of
FGF2 penetrating into injured spinal cord tissue. PEGylation of FGF2 enhanced spinal cord
tissue penetration by nearly twofold when delivered locally from HAMC.
2. Determine the extent of in vivo neuroprotection achieved following localized delivery of
neuroprotective agents to the injured spinal cord
HAMC delivery of EPO, a neuroprotective that crosses the blood brain barrier, showed that a
greater number of motor neurons were spared after SCI relative to intraperitoneal or intrathecal
bolus injections. A reduction of cavity size was also observed, suggesting that localized delivery
of EPO had a tissue sparing effect. FGF2 is a protein that does not cross the intact blood brain
barrier, and its local delivery with HAMC was compared to intratehcal injections of HAMC
alone and aCSF. Relative to these controls, FGF2 delivered with HAMC enhanced spinal cord
blood flow and reduced the permeability-surface area product at sites rostral to the epicenter.
These data suggests that better functional tissue outcomes can be achieved using HAMC as a
local delivery system.
3. Develop a live animal imaging strategy to assess spinal cord hemodynamics after
delivery of neuroprotective agents with HAMC
Both ultrasound paired with laser Doppler and dynamic CT imaging were explored as methods
for assessing SCBF. Dynamic CT imaging required less invasive techniques and was utilized to
investigate SCBF and the permeability-surface area product in the uninjured and injured spinal
cord after localized delivery of FGF2 with HAMC.
Taken together, it is clear that HAMC is an appropriate delivery system for neuroprotective
agents that affect secondary injury events within the first day of injury. Proteins that do or do
not cross the blood spinal cord barrier can be delivered locally with HAMC to mitigate tissue
damage following SCI, demonstrated by neuron sparing achieved with EPO delivery and
improved blood flow and reduced permeability with FGF2 delivery.
105
5.6 Conclusions SCI is a debilitating condition for which there is a lack of safe and effective treatment strategies.
The secondary injury that occurs after the initial mechanical trauma creates a complex
environment that will likely require a combination of pharmacological intervention strategies to
result in a significant amount of functional recovery. Our lab previously developed a polymeric
blend of HAand MC that was assessed in vitro and in vivo to be safe for use as a minimally
invasive intrathecal drug delivery system. This dissertation describes the in vivo potential of
HAMC to locally deliver neuroprotective proteins to the injured spinal cord. Release of
neuroprotective proteins EPO and FGF2 was determined in vitro to occur within the first 16 h
after injection. Intrathecal drug delivery with HAMC maintained protein concentrations on the
surface of the injured spinal cord for several hours, resulting in effective protein concentrations
to diffuse into tissue. Localized delivery of these proteins with HAMC led to neuroprotection
and improved tissue outcomes relative to controls. In addition, dynamic CT imaging was
utilized for the first time to functionally assess SCBF and permeability in live animals, a tool that
could be used as a clinical comparison between animal and human studies of SCI. Taken
together, these results demonstrate that the current formulation of HAMC is an effective and
minimally invasive drug delivery system for delivery of neuroprotective agents to the injured
spinal cord. Its use could be clinically practical for acute SCI treatment following
decompressive surgery performed in the first 24 h after injury, although further work is
necessary to determine the optimal combination of drugs or proteins that will be most effective at
improving functional recovery.
106
6 Limitations and Recommendations for Future Work
6.1 Extending Delivery Window for HAMC Experimental data shown in this dissertation clearly indicate that HAMC has significant potential
as an intrathecal drug delivery system. Neuroprotective agents that can limit the early tissue
impacts of the secondary injury can be effectively delivered to the injured spinal cord with
HAMC. However, one limitation of HAMC is that drug release is largely diffusion-mediated,
thus only short term drug delivery can be achieved for molecules that are simply mixed into the
HAMC blend. Since HAMC only remains in the intrathecal space for 4-7d, conjugation of a
drug or protein to the polymer chains would still generally be limited to delivery during the acute
phase of secondary injury, which can last for up to a month. Additionally, a combinatorial
strategy to maximize the utility of HAMC would include neuroregenerative agents that could
stimulate tissue regrowth after the injury has been resolved. To deliver these molecules from
HAMC would require: 1) reformulation or crosslinking of the hydrogel network to extend the
useful delivery lifetime and 2) a method of controlling the release rate for at least 30 d to deliver
neuroregenerative agents.
6.1.1 Define Mechanism for HAMC Elimination from CNS
To adequately develop HAMC as a drug delivery system, it is important to understand how both
HA and MC are eliminated from the intrathecal space. Both components were initially thought
to dissolve into the CSF and be cleared through the arachnoid granulations, where CSF normally
enters the bloodstream. However, HA was eliminated from the gel more quickly than expected,
suggesting that perhaps it is degraded by free radicals or enzymes that enter through the
hemorrhage following injury. While HA may play a biological role as a therapeutic polymer as
previously discussed, understanding both this role and the mechanism by which it is eliminated
is important for further development of the gel. For example, if it does play an early biological
role, then crosslinking HA to increase the in vivo lifetime would be undesirable. Confirming that
MC is not enzymatically degraded may be important for the same reasons.
107
6.1.2 Extend In vivo Lifetime of HAMC
Increasing the wt% of the polymer in the gel would likely increase the stability of the gel;
however, it also reduces the injectability of the material, an important property for its clinical
relevance. Both HA and MC in the formulation employed in this thesis were of relatively high
molecular weight; however, both polymers are commercially available at higher molecular
weights. Increasing the chain length of individual polymers that are interacting, namely MC,
could increase the lifetime by limiting chain dissolution into CSF. Since each MC polymer unit
possesses several hydrophobic moieties, and hydrophobic association is a dynamic process,
increasing the molecular weight would allow for re-association of some hydrophobic moieties
before the chain becomes disentangled from the polymer network and dissolves away [254].
Also, increasing the chain length of MC makes the path length of diffusion from the gel more
tortuous, thus increasing the time for dissolution of individual polymer chains.
To slow degradation of polymers, chemical crosslinking is commonly used. This could be done
by chemically crosslinking either MC or HA, or by crosslinking HA to MC. This could be done
by chemically modifying the free hydroxyls or carboxyls of HA and MC with reactive moieties
to induce crosslinking, either through a spontaneous mechanism or using an initiator. However,
chemical crosslinking can be toxic when performed in situ. Additionally, chemical crosslinking
often increases the swelling behavior of the material, since chains are no longer able to dissolve
out of the gel and are bound together, producing a large mesh size where water can accumulate.
This property would need to be limited for the intrathecal application so further compression of
the cord does not occur upon injection.
6.1.3 Controlled Release
Extending the release of molecules from polymeric devices can be done using a variety of
techniques. Utilizing strong intermolecular interactions, such as the heparin-FGF2 interaction is
one strategy where release is mediated by dissociation rates of the interacting pairs [319].
Another method could to be attach proteins or drugs to the polymer backbone via enzymatically
cleavable side chains [320], where release could be mediated by the presence of degrading
enzymes during the secondary injury of SCI. A method to significantly increase the length of
108
time that drugs can be delivered is by encapsulating them into solid polymer particles. Release
can be mediated by surface or bulk erosion of the particles and the particles can be synthesized in
varying sizes, even into the nanometer size range. Release can be extended for weeks to months
with polymeric microparticles, potentially allowing for effective neuroregenerative therapies in
SCI patients.
Our lab has pursued some of these avenues and further development of HAMC has shown
promise in vitro for both extended stability and long term delivery [209]. Higher molecular
weight HA and MC in the HAMC blend, paired with poly(lactide-co-glycolide) nanoparticles,
swell similarly to the original formulation of HAMC that was pursued here, degrades in vitro
beyond 56 d. Polymeric particles embedded in the hydrogel can tailor the delivery of molecules
for days to weeks, suggesting that this new composite HAMC can provide sustained and
localized release of therapeutic agents after SCI. In vivo data suggests that it is safe and
biocompatible for this purpose [321].
6.2 Drug Choice Matrix Several neuroprotective agents are being investigated for treatment of SCI, many of which are
summarized in Table 1.3. While HAMC has proved to be a useful material for delivery of two
neuroprotective proteins, significant functional recovery was not observed with either of these
molecules. This could be due to many factors, including dosage delivered, rate of delivery, or
timeframe of delivery. However, due to the complexity of the secondary injury, it is unlikely a
single molecule will be able to mitigate damage in the spinal cord and provide satisfactory
functional recovery for patients. This will likely require a combinatorial strategy, whereby
several of the secondary injury events can be targeted. Although reports for many of these
molecules claim to produce functional recovery, and perhaps statistically better than controls,
this is a phenomenological exploration, not a systematic and mechanistic approach to delivering
an appropriate dose for an adequate amount of time to achieve long term, reproducible functional
recovery. In addition, functional recovery is sometimes not reproducible when the studies are
repeated in a different environment [312, 322]. For this reason, a systematic approach is
proposed here to determine optimal combinations of drugs based on literature support and further
testing.
109
The events of secondary injury and molecules that mediate these events can be categorized into
the following: (1) ion channel antagonists to provide protection from excitotoxicity, (2)
inflammatory and immune regulators, (3) membrane stabilizers and antioxidants, (4) blood flow
mediators and angiogenic agents, (5) antiapoptotic factors and cytoskeletal stabilizers, (6)
neuroregenerative agents, and (7) scar reducing agents. Each drug, which would be categorized
according to these secondary injury events above, could then be assessed based on a set of
weighted criteria to assign a relative value of likelihood that it would be optimal for delivery
with the new formulation of HAMC [209]. For example, each drug should be assessed and
assigned a numerical score based on (1) its ability to cross the BSCB, (2) the extent of reported
systemic side effects, (3) whether clinical trials have been reported for the molecule (in SCI or
otherwise), (4) whether the timeframe of action is clinically relevant, (5) whether it has a
spectrum of mechanistic actions, and (6) the strength of the literature support for its use in SCI.
The score given in each category would then be multiplied by the weighting for each assessment
(based on importance). The sum of these products for each molecule provides a numerical value
that could then be used to rank molecules within a category. An optimal strategy for determining
an effective combination of drugs would be to select drugs from a number of these categories so
that the combination could have a wide range of actions following SCI. This type of assessment
would require an immense amount of literature research to determine an assessment for each
molecule; however, it would also provide an unbiased and systematic approach to determining a
combination of factors that could provide the greatest impact after SCI.
6.3 Development of CT for SC hemodynamics
Dynamic CT imaging has been applied here to measure SCBF and permeability in rats. While
this has advantages over several other techniques for these measurements, further optimization
would be ideal for both practical issues and signal optimization. Dynamic CT imaging is a
minimally invasive and easily applied method in humans because intravenous administration of
contrast agents can be done quite easily. However, intravenous administration is more difficult
in small animals due to the size of vessels that are readily available and go directly to the heart.
In these studies, we attempted to use tail vein administration of contrast agents in rats, however,
successful injections were not possible since the catheters were difficult to insert into the vein
110
and often the catheters did not remain in the vein during the injection of the contrast agent. To
avoid these issues, animals with implanted catheters were used, where contrast was delivered
directly to the heart through catheters inserted into the jugular vein. This is a reproducible and
accepted method of catheterizing small animals; however, it does require a surgical procedure
that is not minimally invasive, and thus not ideal. Additionally, catheters would sometimes
become blocked with clotted blood, at which times the animals could no longer be infused with
contrast for CT imaging. These are significant issues and considerable experience is required to
perform these studies in small animals.
The SCBF measurements obtained were similar to previoiusly published values, suggesting that
dynamic CT is a valid technique. However, many previous studies of SCBF have shown
considerable variability, often correlated to changes in other vital signs, such as mean arterial
blood pressure, hematocrit, and body temperature [53]. In the present study, variability was also
observed, which may have been due to these other physiological paramaters, but since we could
not measure them during CT imaging, they could not be correlated. Permeability measurements
had very high variability, which was more closely correlated to low signal-to-noise ratio and low
sensitivity. These could potentially be improved by varying the x-ray exposure factors, such as
the tube voltage and current. These parameters should be optimized to reduce noise in the signal,
thus allowing more consistent analysis for permeability parameters.
6.4 Limitations for Bench to Bedside Translation
HAMC has proven a useful tool in animal studies for local administration of therapeutic agents,
where relevant concentrations of proteins were achieved within the tissue, and better
neuroprotection was achieved over systemic or intrathecal delivery. However, as described
previously, there are several factors that make it non-ideal for clinical use as of yet.
Reformulation of HAMC and its combination with particles could significantly improve the in
vivo duration and drug delivery capacity; however, these remain to be tested. The formulation
used in these studies would likely not deliver drugs for a sufficiently long period in humans to
offer significant functional benefit, and diffusion distances in humans are also much larger than
those in small animals. Large animal studies to better understand the pharmacokinetics and
111
tissue distribution in vivo relevant to human studies are necessary for utilization in a clinical
setting.
Many small animal studies performed do not truly replicate practical timelines for clinical
studies, one of the potential reasons for many clinical trials failing to show functional
improvement. Animal studies are often performed with drug treatment immediately after SCI.
However, patients may not be …to receive treatment until several hours after injury, thus making
correlation to pre-clinical data difficult. Similarly, HAMC was applied immediately post-injury
in these small animal studies, whereas current recommendations for decompressive surgery in
humans is 8-24 h after injury, the time at which HAMC would optimally be applied clinically.
At this stage in the secondary injury local delivery of many excitotoxic neuroprotectants, such as
NBQX, would likely have little impact since excitotoxicity would have already affected neurons
in a large penumbra of tissue. This suggests that HAMC may be better suited to deliver
neuroprotective agents that attack secondary injury events which occur at least 24 h after the
primary injury. In addition, to make animal studies involving HAMC more clinically relevant,
intrathecal injection should be done 24 h after injury, rather than immediately following injury.
While this is not ideal from the viewpoint of limiting damage after injury, it makes these studies
more likely to indicate the true potential of localized drug delivery with HAMC.
112
References & Bibliography 1. D. Becker, C.L. Sadowsky, and J.W. McDonald, Restoring function after spinal cord
injury, Neurologist. 9 (2003) 1-15
2. C.H. Tator, Strategies for recovery and regeneration after brain and spinal cord injury, Inj Prev. 8 Suppl 4 (2002) IV33-6
3. G.D. Carlson and C. Gorden, Current developments in spinal cord injury research, Spine J. 2 (2002) 116-28
4. N.S.C.I.S. Center. Annual Statistical Report for the Model Spinal Cord Injury Care Systems. 2009 [cited 2009 January 21, 2010]; Public Version:[Available from: https://www.nscisc.uab.edu/public_content/pdf/FactsApr09.pdf.
5. C.D. Reeve Foundation. One Degree of Separation: Paralysis and Spinal Cord Injury in the United States. 2009 [cited 2010 January 14]; Available from: http://www.christopherreeve.org/site/c.mtKZKgMWKwG/b.5184189/k.5587/Paralysis_Facts__Figures.htm.
6. M.G. Fehlings, Editorial: recommendations regarding the use of methylprednisolone in acute spinal cord injury: making sense out of the controversy, Spine. 26 (2001) S56-7
7. R.J. Hurlbert, The role of steroids in acute spinal cord injury: an evidence-based analysis, Spine. 26 (2001) S39-46
8. L.M. Ramer, M.S. Ramer, and J.D. Steeves, Setting the stage for functional repair of spinal cord injuries: a cast of thousands, Spinal Cord. 43 (2005) 134-61
9. M.C. Jimenez Hamann, E.C. Tsai, C.H. Tator, and M.S. Shoichet, Novel intrathecal delivery system for treatment of spinal cord injury, Exp Neurol. 182 (2003) 300-9
10. D. Gupta, C.H. Tator, and M.S. Shoichet, Fast-gelling injectable blend of hyaluronan and methylcellulose for intrathecal, localized delivery to the injured spinal cord, Biomaterials. 27 (2006) 2370-9
11. D. Purves, G.J. Augustine, D. Fitzpatrick, L.C. Katz, A.S. LaMantia, J.O. McNamara, and S.M. Williams, eds. Neuroscience. Second ed. 2001, Sinauer Associates, Inc Publishers: Sunderland, MA.
12. F. Loth, M.A. Yardimci, and N. Alperin, Hydrodynamic modeling of cerebrospinal fluid motion within the spinal cavity, J Biomech Eng. 123 (2001) 71-9
13. J. Fix, Neuroanatomy, 3rd ed. Lippincott Williams & Wilkins, Baltimore, 2002
14. E.R. Kandel, J.H. Schwartz, and T.M. Jessell, eds. Principles of Neural Science. Fourth ed. 2000, McGraw Hill Health Professions Division: New York, NY.
113
15. B.T. Hawkins and T.P. Davis, The blood-brain barrier/neurovascular unit in health and disease, Pharmacol Rev. 57 (2005) 173-85
16. N.J. Abbott, L. Ronnback, and E. Hansson, Astrocyte-endothelial interactions at the blood-brain barrier, Nat Rev Neurosci. 7 (2006) 41-53
17. R.P. Bunge, W.R. Puckett, J.L. Becerra, A. Marcillo, and R.M. Quencer, Observations on the pathology of human spinal cord injury. A review and classification of 22 new cases with details from a case of chronic cord compression with extensive focal demyelination, Adv Neurol. 59 (1993) 75-89
18. M.D. Norenberg, J. Smith, and A. Marcillo, The pathology of human spinal cord injury: defining the problems, J Neurotrauma. 21 (2004) 429-40
19. B.D. Ratner, A.S. Hoffman, F.J. Schoen, and J.E. Lemons, eds. Biomaterials Science: An Introduction to Materials in Medicine. Second ed. 2004, Elsevier Academic Press: San Diego, CA.
20. S.T. DeKosky, S.D. Styren, M.E. O'Malley, J.R. Goss, P. Kochanek, D. Marion, C.H. Evans, and P.D. Robbins, Interleukin-1 receptor antagonist suppresses neurotrophin response in injured rat brain, Ann Neurol. 39 (1996) 123-7
21. M. Schwartz, A. Solomon, V. Lavie, S. Ben-Bassat, M. Belkin, and A. Cohen, Tumor necrosis factor facilitates regeneration of injured central nervous system axons, Brain Res. 545 (1991) 334-8
22. M. Eddleston and L. Mucke, Molecular profile of reactive astrocytes--implications for their role in neurologic disease, Neuroscience. 54 (1993) 15-36
23. Y. Hayashi, M. Nomura, S. Yamagishi, S. Harada, J. Yamashita, and H. Yamamoto, Induction of various blood-brain barrier properties in non-neural endothelial cells by close apposition to co-cultured astrocytes, Glia. 19 (1997) 13-26
24. K. Sobue, N. Yamamoto, K. Yoneda, M.E. Hodgson, K. Yamashiro, N. Tsuruoka, T. Tsuda, H. Katsuya, Y. Miura, K. Asai, and T. Kato, Induction of blood-brain barrier properties in immortalized bovine brain endothelial cells by astrocytic factors, Neurosci Res. 35 (1999) 155-64
25. J.W. Fawcett and R.A. Asher, The glial scar and central nervous system repair, Brain Res Bull. 49 (1999) 377-91
26. D.A. Morgenstern, R.A. Asher, and J.W. Fawcett, Chondroitin sulphate proteoglycans in the CNS injury response, Prog Brain Res. 137 (2002) 313-32
27. C. Rusbridge, D. Greitz, and B.J. Iskandar, Syringomyelia: current concepts in pathogenesis, diagnosis, and treatment, J Vet Intern Med. 20 (2006) 469-79
28. E.D. Hall and J.E. Springer, Neuroprotection and acute spinal cord injury: a reappraisal, NeuroRx. 1 (2004) 80-100
114
29. R.B. Borgens, A.R. Blight, D.J. Murphy, and L. Stewart, Transected dorsal column axons within the guinea pig spinal cord regenerate in the presence of an applied electric field, J Comp Neurol. 250 (1986) 168-80
30. R. Shi and R.B. Borgens, Acute repair of crushed guinea pig spinal cord by polyethylene glycol, J Neurophysiol. 81 (1999) 2406-14
31. P.H. Laverty, A. Leskovar, G.J. Breur, J.R. Coates, R.L. Bergman, W.R. Widmer, J.P. Toombs, S. Shapiro, and R.B. Borgens, A preliminary study of intravenous surfactants in paraplegic dogs: polymer therapy in canine clinical SCI, J Neurotrauma. 21 (2004) 1767-77
32. A.R. Blight, Axonal physiology of chronic spinal cord injury in the cat: intracellular recording in vitro, Neuroscience. 10 (1983) 1471-86
33. D.M. Basso, L.C. Fisher, A.J. Anderson, L.B. Jakeman, D.M. McTigue, and P.G. Popovich, Basso Mouse Scale for locomotion detects differences in recovery after spinal cord injury in five common mouse strains, J Neurotrauma. 23 (2006) 635-59
34. A.N. Sandler and C.H. Tator, Regional spinal cord blood flow in primates, J Neurosurg. 45 (1976) 647-59
35. C.H. Tator, Experimental circumferential compression injury of primate spinal cord, Proc Veterans Adm Spinal Cord Inj Conf. 18 (1971) 2-5
36. J.C. Fleming, M.D. Norenberg, D.A. Ramsay, G.A. Dekaban, A.E. Marcillo, A.D. Saenz, M. Pasquale-Styles, W.D. Dietrich, and L.C. Weaver, The cellular inflammatory response in human spinal cords after injury, Brain. 129 (2006) 3249-69
37. D.M. McTigue, P.G. Popovich, L.B. Jakeman, and B.T. Stokes, Strategies for spinal cord injury repair, Prog Brain Res. 128 (2000) 3-8
38. G.A. Metz, A. Curt, H. van de Meent, I. Klusman, M.E. Schwab, and V. Dietz, Validation of the weight-drop contusion model in rats: a comparative study of human spinal cord injury, J Neurotrauma. 17 (2000) 1-17
39. L. Guth, D. Bright, and E.J. Donati, Functional deficits and anatomical alterations after high cervical spinal hemisection in the rat, Exp Neurol. 58 (1978) 511-20
40. C. Scheifer, U. Hoheisel, P. Trudrung, T. Unger, and S. Mense, Rats with chronic spinal cord transection as a possible model for the at-level pain of paraplegic patients, Neurosci Lett. 323 (2002) 117-20
41. E.J. Dolan, C.H. Tator, and L. Endrenyi, The value of decompression for acute experimental spinal cord compression injury, J Neurosurg. 53 (1980) 749-55
42. E.J. Dolan and C.H. Tator, A new method for testing the force of clips for aneurysms or experimental spinal cord compression, J Neurosurg. 51 (1979) 229-33
115
43. A.S. Rivlin and C.H. Tator, Effect of duration of acute spinal cord compression in a new acute cord injury model in the rat, Surg Neurol. 10 (1978) 38-43
44. J.A. Gruner, A monitored contusion model of spinal cord injury in the rat, J Neurotrauma. 9 (1992) 123-6; discussion 126-8
45. L.B. Jakeman, Z. Guan, P. Wei, R. Ponnappan, R. Dzwonczyk, P.G. Popovich, and B.T. Stokes, Traumatic spinal cord injury produced by controlled contusion in mouse, J Neurotrauma. 17 (2000) 299-319
46. S.M. Onifer, A.G. Rabchevsky, and S.W. Scheff, Rat models of traumatic spinal cord injury to assess motor recovery, Ilar J. 48 (2007) 385-95
47. R. Talac, J.A. Friedman, M.J. Moore, L. Lu, E. Jabbari, A.J. Windebank, B.L. Currier, and M.J. Yaszemski, Animal models of spinal cord injury for evaluation of tissue engineering treatment strategies, Biomaterials. 25 (2004) 1505-10
48. E.S. Rosenzweig and J.W. McDonald, Rodent models for treatment of spinal cord injury: research trends and progress toward useful repair, Curr Opin Neurol. 17 (2004) 121-31
49. A. Gorio, N. Gokmen, S. Erbayraktar, O. Yilmaz, L. Madaschi, C. Cichetti, A.M. Di Giulio, E. Vardar, A. Cerami, and M. Brines, Recombinant human erythropoietin counteracts secondary injury and markedly enhances neurological recovery from experimental spinal cord trauma, Proc Natl Acad Sci U S A. 99 (2002) 9450-5
50. A.N. Sandler and C.H. Tator, Effect of acute spinal cord compression injury on regional spinal cord blood flow in primates, J Neurosurg. 45 (1976) 660-76
51. P.C. Poon, D. Gupta, M.S. Shoichet, and C.H. Tator, Clip compression model is useful for thoracic spinal cord injuries: histologic and functional correlates, Spine. 32 (2007) 2853-9
52. L.J. Noble and J.R. Wrathall, Distribution and time course of protein extravasation in the rat spinal cord after contusive injury, Brain Res. 482 (1989) 57-66
53. A. Guha, C.H. Tator, and J. Rochon, Spinal cord blood flow and systemic blood pressure after experimental spinal cord injury in rats, Stroke. 20 (1989) 372-7
54. A.R. Blight, Morphometric analysis of a model of spinal cord injury in guinea pigs, with behavioral evidence of delayed secondary pathology, J Neurol Sci. 103 (1991) 156-71
55. P.G. Popovich, P.J. Horner, B.B. Mullin, and B.T. Stokes, A quantitative spatial analysis of the blood-spinal cord barrier. I. Permeability changes after experimental spinal contusion injury, Exp Neurol. 142 (1996) 258-75
56. M. Bilgen, B. Dogan, and P.A. Narayana, In vivo assessment of blood-spinal cord barrier permeability: serial dynamic contrast enhanced MRI of spinal cord injury, Magn Reson Imaging. 20 (2002) 337-41
116
57. M. Bilgen, R. Abbe, and P.A. Narayana, Dynamic contrast-enhanced MRI of experimental spinal cord injury: in vivo serial studies, Magn Reson Med. 45 (2001) 614-22
58. S.S. Kety, The theory and applications of the exchange of inert gas at the lungs and tissues, Pharmacol Rev. 3 (1951) 1-41
59. S.S. Kety and C.F. Schmidt, The nitrous oxide method for the quantitative determination of cerebral blood flow in man; theory, procedure and normal values, J Clin Invest. 27 (1948) 476-83
60. I.R. Griffiths, Spinal cord blood flow in dogs. 1. The 'normal' flow, J Neurol Neurosurg Psychiatry. 36 (1973) 34-41
61. A.L. Smith, J.W. Pender, and S.C. Alexander, Effects of PCO2 on spinal cord blood flow, Am J Physiol. 216 (1969) 1158-63
62. A.I. Kobrine, T.F. Doyle, and A.N. Martins, Spinal cord blood flow in the rhesus monkey by the hydrogen clearance method, Surg Neurol. 2 (1974) 197-200
63. W.M. Landau, W.H. Freygang, Jr., L.P. Roland, L. Sokoloff, and S.S. Kety, The local circulation of the living brain; values in the unanesthetized and anesthetized cat, Trans Am Neurol Assoc. (1955) 125-9
64. M. Reivich, J. Jehle, L. Sokoloff, and S.S. Kety, Measurement of regional cerebral blood flow with antipyrine-14C in awake cats, J Appl Physiol. 27 (1969) 296-300
65. Y. Malik, D. Spreng, M. Konar, M.G. Doherr, A. Jaggy, J. Howard, and F. Forterre, Laser-Doppler measurements of spinal cord blood flow changes during hemilaminectomy in chondrodystrophic dogs with disk extrusion, Vet Surg. 38 (2009) 457-62
66. S. Bisdas, Z. Rumboldt, K. Surlan, T.S. Koh, J. Deveikis, and M.V. Spampinato, Perfusion CT measurements in healthy cervical spinal cord: feasibility and repeatability of the study as well as interchangeability of the perfusion estimates using two commercially available software packages, Eur Radiol. 18 (2008) 2321-8
67. D.A. Leggett, K.A. Miles, and B.B. Kelley, Blood-brain barrier and blood volume imaging of cerebral glioma using functional CT: a pictorial review, Australas Radiol. 42 (1998) 335-40
68. K.A. Miles, Measurement of tissue perfusion by dynamic computed tomography, Br J Radiol. 64 (1991) 409-12
69. K.A. Miles, C. Charnsangavej, F.T. Lee, E.K. Fishman, K. Horton, and T.Y. Lee, Application of CT in the investigation of angiogenesis in oncology, Acad Radiol. 7 (2000) 840-50
70. K.A. Miles, M.P. Hayball, and A.K. Dixon, Functional images of hepatic perfusion obtained with dynamic CT, Radiology. 188 (1993) 405-11
117
71. K.A. Miles, D.A. Leggett, B.B. Kelley, M.P. Hayball, R. Sinnatamby, and I. Bunce, In vivo assessment of neovascularization of liver metastases using perfusion CT, Br J Radiol. 71 (1998) 276-81
72. U. Tateishi, H. Nishihara, S. Watanabe, T. Morikawa, K. Abe, and K. Miyasaka, Tumor angiogenesis and dynamic CT in lung adenocarcinoma: radiologic-pathologic correlation, J Comput Assist Tomogr. 25 (2001) 23-7
73. W.T. Yeung, T.Y. Lee, R.F. Del Maestro, R. Kozak, J. Bennett, and T. Brown, Effect of steroids on iopamidol blood-brain transfer constant and plasma volume in brain tumors measured with X-ray computed tomography, J Neurooncol. 18 (1994) 53-60
74. C.A. Yi, K.S. Lee, E.A. Kim, J. Han, H. Kim, O.J. Kwon, Y.J. Jeong, and S. Kim, Solitary pulmonary nodules: dynamic enhanced multi-detector row CT study and comparison with vascular endothelial growth factor and microvessel density, Radiology. 233 (2004) 191-9
75. P.S. Tofts, G. Brix, D.L. Buckley, J.L. Evelhoch, E. Henderson, M.V. Knopp, H.B. Larsson, T.Y. Lee, N.A. Mayr, G.J. Parker, R.E. Port, J. Taylor, and R.M. Weisskoff, Estimating kinetic parameters from dynamic contrast-enhanced T(1)-weighted MRI of a diffusable tracer: standardized quantities and symbols, J Magn Reson Imaging. 10 (1999) 223-32
76. P.S. Tofts, Modeling tracer kinetics in dynamic Gd-DTPA MR imaging, J Magn Reson Imaging. 7 (1997) 91-101
77. K.S. St Lawrence and T.Y. Lee, An adiabatic approximation to the tissue homogeneity model for water exchange in the brain: II. Experimental validation, J Cereb Blood Flow Metab. 18 (1998) 1378-85
78. K.S. St Lawrence and T.Y. Lee, An adiabatic approximation to the tissue homogeneity model for water exchange in the brain: I. Theoretical derivation, J Cereb Blood Flow Metab. 18 (1998) 1365-77
79. A.S. Rivlin and C.H. Tator, Objective clinical assessment of motor function after experimental spinal cord injury in the rat, J Neurosurg. 47 (1977) 577-81
80. S.P. Hicks and C.J. D'Amato, Motor-sensory cortex-corticospinal system and developing locomotion and placing in rats, Am J Anat. 143 (1975) 1-42
81. G.A. Metz and I.Q. Whishaw, Cortical and subcortical lesions impair skilled walking in the ladder rung walking test: a new task to evaluate fore- and hindlimb stepping, placing, and co-ordination, J Neurosci Methods. 115 (2002) 169-79
82. W.J. Z'Graggen, G.A. Metz, G.L. Kartje, M. Thallmair, and M.E. Schwab, Functional recovery and enhanced corticofugal plasticity after unilateral pyramidal tract lesion and blockade of myelin-associated neurite growth inhibitors in adult rats, J Neurosci. 18 (1998) 4744-57
118
83. L. de Medinaceli, W.J. Freed, and R.J. Wyatt, An index of the functional condition of rat sciatic nerve based on measurements made from walking tracks, Exp Neurol. 77 (1982) 634-43
84. D.M. Basso, M.S. Beattie, and J.C. Bresnahan, A sensitive and reliable locomotor rating scale for open field testing in rats, J Neurotrauma. 12 (1995) 1-21
85. G.A. Metz, D. Merkler, V. Dietz, M.E. Schwab, and K. Fouad, Efficient testing of motor function in spinal cord injured rats, Brain Res. 883 (2000) 165-77
86. D.M. Basso, M.S. Beattie, J.C. Bresnahan, D.K. Anderson, A.I. Faden, J.A. Gruner, T.R. Holford, C.Y. Hsu, L.J. Noble, R. Nockels, P.L. Perot, S.K. Salzman, and W. Young, MASCIS evaluation of open field locomotor scores: effects of experience and teamwork on reliability. Multicenter Animal Spinal Cord Injury Study, J Neurotrauma. 13 (1996) 343-59
87. M.C. Jimenez Hamann, C.H. Tator, and M.S. Shoichet, Injectable intrathecal delivery system for localized administration of EGF and FGF-2 to the injured rat spinal cord, Exp Neurol. 194 (2005) 106-19
88. D.D. Pearse and D.J. Barakat, Cellular repair strategies for spinal cord injury, Expert Opin Biol Ther. 6 (2006) 639-52
89. I. Kulbatski, A.J. Mothe, H. Nomura, and C.H. Tator, Endogenous and exogenous CNS derived stem/progenitor cell approaches for neurotrauma, Curr Drug Targets. 6 (2005) 111-26
90. H. Nomura, C.H. Tator, and M.S. Shoichet, Bioengineered strategies for spinal cord repair, J Neurotrauma. 23 (2006) 496-507
91. D. Wu, B.W. Song, H.V. Vinters, and W.M. Pardridge, Pharmacokinetics and brain uptake of biotinylated basic fibroblast growth factor conjugated to a blood-brain barrier drug delivery system, J Drug Target. 10 (2002) 239-45
92. D. Wu and W.M. Pardridge, Neuroprotection with noninvasive neurotrophin delivery to the brain, Proc Natl Acad Sci U S A. 96 (1999) 254-9
93. E.C. Tsai and C.H. Tator, Neuroprotection and regeneration strategies for spinal cord repair, Curr Pharm Des. 11 (2005) 1211-22
94. E.C. Tsai, P.D. Dalton, M.S. Shoichet, and C.H. Tator, Synthetic hydrogel guidance channels facilitate regeneration of adult rat brainstem motor axons after complete spinal cord transection, J Neurotrauma. 21 (2004) 789-804
95. H. Nomura, Y. Katayama, M.S. Shoichet, and C.H. Tator, Complete spinal cord transection treated by implantation of a reinforced synthetic hydrogel channel results in syringomyelia and caudal migration of the rostral stump, Neurosurgery. 59 (2006) 183-92; discussion 183-92
119
96. P.J. Reier, Cellular transplantation strategies for spinal cord injury and translational neurobiology, NeuroRx. 1 (2004) 424-51
97. M.G. Fehlings and C.H. Tator, An evidence-based review of decompressive surgery in acute spinal cord injury: rationale, indications, and timing based on experimental and clinical studies, J Neurosurg. 91 (1999) 1-11
98. J.S. Silber and A.R. Vaccaro, Summary statement: the role and timing of decompression in acute spinal cord injury: evidence-based guidelines, Spine (Phila Pa 1976). 26 (2001) S110
99. J.C. Furlan, V. Noonan, D.W. Cadotte, and M. Fehlings, Timing of decompressive surgery after traumatic spinal cord injury: An evidence-based examination of pre-clinical and clinical studies, J Neurotrauma. In Press (2009
100. J. Xu, G. Fan, S. Chen, Y. Wu, X.M. Xu, and C.Y. Hsu, Methylprednisolone inhibition of TNF-alpha expression and NF-kB activation after spinal cord injury in rats, Brain Res Mol Brain Res. 59 (1998) 135-42
101. M.B. Bracken, W.F. Collins, D.F. Freeman, M.J. Shepard, F.W. Wagner, R.M. Silten, K.G. Hellenbrand, J. Ransohoff, W.E. Hunt, P.L. Perot, Jr., and et al., Efficacy of methylprednisolone in acute spinal cord injury, Jama. 251 (1984) 45-52
102. M.B. Bracken, M.J. Shepard, K.G. Hellenbrand, W.F. Collins, L.S. Leo, D.F. Freeman, F.C. Wagner, E.S. Flamm, H.M. Eisenberg, J.H. Goodman, and et al., Methylprednisolone and neurological function 1 year after spinal cord injury. Results of the National Acute Spinal Cord Injury Study, J Neurosurg. 63 (1985) 704-13
103. M.B. Bracken, M.J. Shepard, W.F. Collins, T.R. Holford, W. Young, D.S. Baskin, H.M. Eisenberg, E. Flamm, L. Leo-Summers, J. Maroon, and et al., A randomized, controlled trial of methylprednisolone or naloxone in the treatment of acute spinal-cord injury. Results of the Second National Acute Spinal Cord Injury Study, N Engl J Med. 322 (1990) 1405-11
104. M.B. Bracken, M.J. Shepard, W.F. Collins, Jr., T.R. Holford, D.S. Baskin, H.M. Eisenberg, E. Flamm, L. Leo-Summers, J.C. Maroon, L.F. Marshall, and et al., Methylprednisolone or naloxone treatment after acute spinal cord injury: 1-year follow-up data. Results of the second National Acute Spinal Cord Injury Study, J Neurosurg. 76 (1992) 23-31
105. M.B. Bracken, M.J. Shepard, T.R. Holford, L. Leo-Summers, E.F. Aldrich, M. Fazl, M. Fehlings, D.L. Herr, P.W. Hitchon, L.F. Marshall, R.P. Nockels, V. Pascale, P.L. Perot, Jr., J. Piepmeier, V.K. Sonntag, F. Wagner, J.E. Wilberger, H.R. Winn, and W. Young, Administration of methylprednisolone for 24 or 48 hours or tirilazad mesylate for 48 hours in the treatment of acute spinal cord injury. Results of the Third National Acute Spinal Cord Injury Randomized Controlled Trial. National Acute Spinal Cord Injury Study, Jama. 277 (1997) 1597-604
120
106. L.H. Pitts, A. Ross, G.A. Chase, and A.I. Faden, Treatment with thyrotropin-releasing hormone (TRH) in patients with traumatic spinal cord injuries, J Neurotrauma. 12 (1995) 235-43
107. F.H. Geisler, F.C. Dorsey, and W.P. Coleman, Recovery of motor function after spinal-cord injury--a randomized, placebo-controlled trial with GM-1 ganglioside, N Engl J Med. 324 (1991) 1829-38
108. F.H. Geisler, F.C. Dorsey, and W.P. Coleman, GM-1 ganglioside in human spinal cord injury, J Neurotrauma. 9 Suppl 1 (1992) S407-16
109. J.F. Lepeintre, P. D'Arbigny, J.F. Mathe, B. Vigue, G. Loubert, J. Delcour, C. Kempf, and M. Tadie, Neuroprotective effect of gacyclidine. A multicenter double-blind pilot trial in patients with acute traumatic brain injury, Neurochirurgie. 50 (2004) 83-95
110. M.E. Petitjean, V. Pointillart, F. Dixmerias, L. Wiart, F. Sztark, P. Lassie, M. Thicoipe, and P. Dabadie, [Medical treatment of spinal cord injury in the acute stage], Ann Fr Anesth Reanim. 17 (1998) 114-22
111. J.B. Long, A. Martinez-Arizala, J.M. Petras, and J.W. Holaday, Endogenous opioids in spinal cord injury: a critical evaluation, Cent Nerv Syst Trauma. 3 (1986) 295-315
112. S.S. Haghighi, E.D. Hall, X.Z. Geng, J.J. Oro, and G.C. Johnson, Therapeutic value of 21-aminosteroid U74389F in acute spinal cord injury, Neurol Res. 15 (1993) 321-6
113. E.D. Hall, Lipid antioxidants in acute central nervous system injury, Ann Emerg Med. 22 (1993) 1022-7
114. G.W. Hawryluk, J. Rowland, B.K. Kwon, and M.G. Fehlings, Protection and repair of the injured spinal cord: a review of completed, ongoing, and planned clinical trials for acute spinal cord injury, Neurosurg Focus. 25 (2008) E14
115. L.L. Jones and M.H. Tuszynski, Chronic intrathecal infusions after spinal cord injury cause scarring and compression, Microsc Res Tech. 54 (2001) 317-24
116. G. Onose, A. Anghelescu, D.F. Muresanu, L. Padure, M.A. Haras, C.O. Chendreanu, L.V. Onose, A. Mirea, A.V. Ciurea, W.S. El Masri, and K.R. von Wild, A review of published reports on neuroprotection in spinal cord injury, Spinal Cord. 47 (2009) 716-26
117. M.G. Fehlings, C.H. Tator, and R.D. Linden, The effect of nimodipine and dextran on axonal function and blood flow following experimental spinal cord injury, J Neurosurg. 71 (1989) 403-16
118. J.R. Wrathall, Y.D. Teng, and R. Marriott, Delayed antagonism of AMPA/kainate receptors reduces long-term functional deficits resulting from spinal cord trauma, Exp Neurol. 145 (1997) 565-73
121
119. M. Gaviria, A. Privat, P. d'Arbigny, J. Kamenka, H. Haton, and F. Ohanna, Neuroprotective effects of a novel NMDA antagonist, Gacyclidine, after experimental contusive spinal cord injury in adult rats, Brain Res. 874 (2000) 200-9
120. M. Gaviria, A. Privat, P. d'Arbigny, J.M. Kamenka, H. Haton, and F. Ohanna, Neuroprotective effects of gacyclidine after experimental photochemical spinal cord lesion in adult rats: dose-window and time-window effects, J Neurotrauma. 17 (2000) 19-30
121. R.C. Gupta, Y. Seki, and J. Yosida, Role of taurine in spinal cord injury, Curr Neurovasc Res. 3 (2006) 225-35
122. C.D. Mills, K.M. Johnson, and C.E. Hulsebosch, Group I metabotropic glutamate receptors in spinal cord injury: roles in neuroprotection and the development of chronic central pain, J Neurotrauma. 19 (2002) 23-42
123. L.J. Rosenberg, Y.D. Teng, and J.R. Wrathall, Effects of the sodium channel blocker tetrodotoxin on acute white matter pathology after experimental contusive spinal cord injury, J Neurosci. 19 (1999) 6122-33
124. S.J. Wang, K.Y. Wang, and W.C. Wang, Mechanisms underlying the riluzole inhibition of glutamate release from rat cerebral cortex nerve terminals (synaptosomes), Neuroscience. 125 (2004) 191-201
125. G. Schwartz and M.G. Fehlings, Evaluation of the neuroprotective effects of sodium channel blockers after spinal cord injury: improved behavioral and neuroanatomical recovery with riluzole, J Neurosurg. 94 (2001) 245-56
126. P.A. Schumacher, J.H. Eubanks, and M.G. Fehlings, Increased calpain I-mediated proteolysis, and preferential loss of dephosphorylated NF200, following traumatic spinal cord injury, Neuroscience. 91 (1999) 733-44
127. P.A. Schumacher, R.G. Siman, and M.G. Fehlings, Pretreatment with calpain inhibitor CEP-4143 inhibits calpain I activation and cytoskeletal degradation, improves neurological function, and enhances axonal survival after traumatic spinal cord injury, J Neurochem. 74 (2000) 1646-55
128. A. Ibarra and A. Diaz-Ruiz, Protective effect of cyclosporin-A in spinal cord injury: an overview, Curr Med Chem. 13 (2006) 2703-10
129. P.J. Mabon, L.C. Weaver, and G.A. Dekaban, Inhibition of monocyte/macrophage migration to a spinal cord injury site by an antibody to the integrin alphaD: a potential new anti-inflammatory treatment, Exp Neurol. 166 (2000) 52-64
130. Y. Taoka, K. Okajima, M. Uchiba, K. Murakami, S. Kushimoto, M. Johno, M. Naruo, H. Okabe, and K. Takatsuki, Role of neutrophils in spinal cord injury in the rat, Neuroscience. 79 (1997) 1177-82
122
131. J.E. Wells, R.J. Hurlbert, M.G. Fehlings, and V.W. Yong, Neuroprotection by minocycline facilitates significant recovery from spinal cord injury in mice, Brain. 126 (2003) 1628-37
132. M. Thomas, W.D. Le, and J. Jankovic, Minocycline and other tetracycline derivatives: a neuroprotective strategy in Parkinson's disease and Huntington's disease, Clin Neuropharmacol. 26 (2003) 18-23
133. J. Luo, R. Borgens, and R. Shi, Polyethylene glycol immediately repairs neuronal membranes and inhibits free radical production after acute spinal cord injury, J Neurochem. 83 (2002) 471-80
134. J. Donaldson, R. Shi, and R. Borgens, Polyethylene glycol rapidly restores physiological functions in damaged sciatic nerves of guinea pigs, Neurosurgery. 50 (2002) 147-56; discussion 156-7
135. R. Schweigreiter and C.E. Bandtlow, Nogo in the injured spinal cord, J Neurotrauma. 23 (2006) 384-96
136. A.B. Huber and M.E. Schwab, Nogo-A, a potent inhibitor of neurite outgrowth and regeneration, Biol Chem. 381 (2000) 407-19
137. T. Oertle, M.E. van der Haar, C.E. Bandtlow, A. Robeva, P. Burfeind, A. Buss, A.B. Huber, M. Simonen, L. Schnell, C. Brosamle, K. Kaupmann, R. Vallon, and M.E. Schwab, Nogo-A inhibits neurite outgrowth and cell spreading with three discrete regions, J Neurosci. 23 (2003) 5393-406
138. A.G. Rabchevsky, I. Fugaccia, A. Fletcher-Turner, D.A. Blades, M.P. Mattson, and S.W. Scheff, Basic fibroblast growth factor (bFGF) enhances tissue sparing and functional recovery following moderate spinal cord injury, J Neurotrauma. 16 (1999) 817-30
139. A.G. Rabchevsky, I. Fugaccia, A.F. Turner, D.A. Blades, M.P. Mattson, and S.W. Scheff, Basic fibroblast growth factor (bFGF) enhances functional recovery following severe spinal cord injury to the rat, Exp Neurol. 164 (2000) 280-91
140. R. Vavrek, J. Girgis, W. Tetzlaff, G.W. Hiebert, and K. Fouad, BDNF promotes connections of corticospinal neurons onto spared descending interneurons in spinal cord injured rats, Brain. 129 (2006) 1534-45
141. A. Kojima and C.H. Tator, Epidermal growth factor and fibroblast growth factor 2 cause proliferation of ependymal precursor cells in the adult rat spinal cord in vivo, J Neuropathol Exp Neurol. 59 (2000) 687-97
142. E. Kaptanoglu, I. Solaroglu, O. Okutan, H.S. Surucu, F. Akbiyik, and E. Beskonakli, Erythropoietin exerts neuroprotection after acute spinal cord injury in rats: effect on lipid peroxidation and early ultrastructural findings, Neurosurg Rev. 27 (2004) 113-20
123
143. S.J. Taylor, J.W. McDonald, 3rd, and S.E. Sakiyama-Elbert, Controlled release of neurotrophin-3 from fibrin gels for spinal cord injury, J Control Release. 98 (2004) 281-94
144. S.J. Taylor, E.S. Rosenzweig, J.W. McDonald, 3rd, and S.E. Sakiyama-Elbert, Delivery of neurotrophin-3 from fibrin enhances neuronal fiber sprouting after spinal cord injury, J Control Release. 113 (2006) 226-35
145. R.J. Dumont, D.O. Okonkwo, S. Verma, R.J. Hurlbert, P.T. Boulos, D.B. Ellegala, and A.S. Dumont, Acute spinal cord injury, part I: pathophysiologic mechanisms, Clin Neuropharmacol. 24 (2001) 254-64
146. M. Li, V.O. Ona, M. Chen, M. Kaul, L. Tenneti, X. Zhang, P.E. Stieg, S.A. Lipton, and R.M. Friedlander, Functional role and therapeutic implications of neuronal caspase-1 and -3 in a mouse model of traumatic spinal cord injury, Neuroscience. 99 (2000) 333-42
147. J.E. Springer, R.D. Azbill, and P.E. Knapp, Activation of the caspase-3 apoptotic cascade in traumatic spinal cord injury, Nat Med. 5 (1999) 943-6
148. P. Dergham, B. Ellezam, C. Essagian, H. Avedissian, W.D. Lubell, and L. McKerracher, Rho signaling pathway targeted to promote spinal cord repair, J Neurosci. 22 (2002) 6570-7
149. U. Laufs, M. Endres, N. Stagliano, S. Amin-Hanjani, D.S. Chui, S.X. Yang, T. Simoncini, M. Yamada, E. Rabkin, P.G. Allen, P.L. Huang, M. Bohm, F.J. Schoen, M.A. Moskowitz, and J.K. Liao, Neuroprotection mediated by changes in the endothelial actin cytoskeleton, J Clin Invest. 106 (2000) 15-24
150. D.C. Baptiste and M.G. Fehlings, Pharmacological approaches to repair the injured spinal cord, J Neurotrauma. 23 (2006) 318-34
151. S.E. Graber and S.B. Krantz, Erythropoietin and the control of red cell production, Annu Rev Med. 29 (1978) 51-66
152. J. Lah, I. Prislan, B. Krzan, M. Salobir, A. Francky, and G. Vesnaver, Erythropoietin unfolding: thermodynamics and its correlation with structural features, Biochemistry. 44 (2005) 13883-92
153. M. Digicaylioglu, S. Bichet, H.H. Marti, R.H. Wenger, L.A. Rivas, C. Bauer, and M. Gassmann, Localization of specific erythropoietin binding sites in defined areas of the mouse brain, Proc Natl Acad Sci U S A. 92 (1995) 3717-20
154. H.H. Marti, R.H. Wenger, L.A. Rivas, U. Straumann, M. Digicaylioglu, V. Henn, Y. Yonekawa, C. Bauer, and M. Gassmann, Erythropoietin gene expression in human, monkey and murine brain, Eur J Neurosci. 8 (1996) 666-76
155. E. Morishita, S. Masuda, M. Nagao, Y. Yasuda, and R. Sasaki, Erythropoietin receptor is expressed in rat hippocampal and cerebral cortical neurons, and erythropoietin prevents in vitro glutamate-induced neuronal death, Neuroscience. 76 (1997) 105-16
124
156. M. Bernaudin, H.H. Marti, S. Roussel, D. Divoux, A. Nouvelot, E.T. MacKenzie, and E. Petit, A potential role for erythropoietin in focal permanent cerebral ischemia in mice, J Cereb Blood Flow Metab. 19 (1999) 643-51
157. A.L. Siren, F. Knerlich, W. Poser, C.H. Gleiter, W. Bruck, and H. Ehrenreich, Erythropoietin and erythropoietin receptor in human ischemic/hypoxic brain, Acta Neuropathol (Berl). 101 (2001) 271-6
158. A. Kretz, C.J. Happold, J.K. Marticke, and S. Isenmann, Erythropoietin promotes regeneration of adult CNS neurons via Jak2/Stat3 and PI3K/AKT pathway activation, Mol Cell Neurosci. 29 (2005) 569-79
159. Y. Konishi, D.H. Chui, H. Hirose, T. Kunishita, and T. Tabira, Trophic effect of erythropoietin and other hematopoietic factors on central cholinergic neurons in vitro and in vivo, Brain Res. 609 (1993) 29-35
160. P. Lewczuk, M. Hasselblatt, H. Kamrowski-Kruck, A. Heyer, C. Unzicker, A.L. Siren, and H. Ehrenreich, Survival of hippocampal neurons in culture upon hypoxia: effect of erythropoietin, Neuroreport. 11 (2000) 3485-8
161. M. Celik, N. Gokmen, S. Erbayraktar, M. Akhisaroglu, S. Konakc, C. Ulukus, S. Genc, K. Genc, E. Sagiroglu, A. Cerami, and M. Brines, Erythropoietin prevents motor neuron apoptosis and neurologic disability in experimental spinal cord ischemic injury, Proc Natl Acad Sci U S A. 99 (2002) 2258-63
162. D. Agnello, P. Bigini, P. Villa, T. Mennini, A. Cerami, M.L. Brines, and P. Ghezzi, Erythropoietin exerts an anti-inflammatory effect on the CNS in a model of experimental autoimmune encephalomyelitis, Brain Res. 952 (2002) 128-34
163. P. Villa, P. Bigini, T. Mennini, D. Agnello, T. Laragione, A. Cagnotto, B. Viviani, M. Marinovich, A. Cerami, T.R. Coleman, M. Brines, and P. Ghezzi, Erythropoietin selectively attenuates cytokine production and inflammation in cerebral ischemia by targeting neuronal apoptosis, J Exp Med. 198 (2003) 971-5
164. Z. Diaz, M.I. Assaraf, W.H. Miller, Jr., and H.M. Schipper, Astroglial cytoprotection by erythropoietin pre-conditioning: implications for ischemic and degenerative CNS disorders, J Neurochem. 93 (2005) 392-402
165. M. Sugawa, Y. Sakurai, Y. Ishikawa-Ieda, H. Suzuki, and H. Asou, Effects of erythropoietin on glial cell development; oligodendrocyte maturation and astrocyte proliferation, Neurosci Res. 44 (2002) 391-403
166. S. Masuda, M. Okano, K. Yamagishi, M. Nagao, M. Ueda, and R. Sasaki, A novel site of erythropoietin production. Oxygen-dependent production in cultured rat astrocytes, J Biol Chem. 269 (1994) 19488-93
167. E.J. Demers, R.J. McPherson, and S.E. Juul, Erythropoietin protects dopaminergic neurons and improves neurobehavioral outcomes in juvenile rats after neonatal hypoxia-ischemia, Pediatr Res. 58 (2005) 297-301
125
168. A. Aydin, K. Genc, M. Akhisaroglu, K. Yorukoglu, N. Gokmen, and E. Gonullu, Erythropoietin exerts neuroprotective effect in neonatal rat model of hypoxic-ischemic brain injury, Brain Dev. 25 (2003) 494-8
169. H. Ehrenreich, M. Hasselblatt, C. Dembowski, L. Cepek, P. Lewczuk, M. Stiefel, H.H. Rustenbeck, N. Breiter, S. Jacob, F. Knerlich, M. Bohn, W. Poser, E. Ruther, M. Kochen, O. Gefeller, C. Gleiter, T.C. Wessel, M. De Ryck, L. Itri, H. Prange, A. Cerami, M. Brines, and A.L. Siren, Erythropoietin therapy for acute stroke is both safe and beneficial, Mol Med. 8 (2002) 495-505
170. G. Grasso, A. Sfacteria, F. Meli, V. Fodale, M. Buemi, and D.G. Iacopino, Neuroprotection by erythropoietin administration after experimental traumatic brain injury, Brain Res. 1182 (2007) 99-105
171. A. Cetin, K. Nas, H. Buyukbayram, A. Ceviz, and G. Olmez, The effects of systemically administered methylprednisolone and recombinant human erythropoietin after acute spinal cord compressive injury in rats, Eur Spine J. 15 (2006) 1539-44
172. M.L. Brines, P. Ghezzi, S. Keenan, D. Agnello, N.C. de Lanerolle, C. Cerami, L.M. Itri, and A. Cerami, Erythropoietin crosses the blood-brain barrier to protect against experimental brain injury, Proc Natl Acad Sci U S A. 97 (2000) 10526-31
173. A.M. Gonzalez, L.S. Carman, M. Ong, J. Ray, F.H. Gage, C.W. Shults, and A. Baird, Storage, metabolism, and processing of 125I-fibroblast growth factor-2 after intracerebral injection, Brain Res. 665 (1994) 285-92
174. M.A. Nugent and R.V. Iozzo, Fibroblast growth factor-2, Int J Biochem Cell Biol. 32 (2000) 115-20
175. F.J. Moy, A.P. Seddon, P. Bohlen, and R. Powers, High-resolution solution structure of basic fibroblast growth factor determined by multidimensional heteronuclear magnetic resonance spectroscopy, Biochemistry. 35 (1996) 13552-61
176. J.D. Zhang, L.S. Cousens, P.J. Barr, and S.R. Sprang, 3-Dimensional Structure of Human Basic Fibroblast Growth-Factor, a Structural Homolog of Interleukin-1-Beta, Proceedings of the National Academy of Sciences of the United States of America. 88 (1991) 3446-3450
177. S.J. Prestrelski, G.M. Fox, and T. Arakawa, Binding of heparin to basic fibroblast growth factor induces a conformational change, Arch Biochem Biophys. 293 (1992) 314-9
178. S. Faham, R.E. Hileman, J.R. Fromm, R.J. Linhardt, and D.C. Rees, Heparin structure and interactions with basic fibroblast growth factor, Science. 271 (1996) 1116-20
179. D. Gospodarowicz and J. Cheng, Heparin protects basic and acidic FGF from inactivation, J Cell Physiol. 128 (1986) 475-84
126
180. Y. Shing, J. Folkman, C. Haudenschild, D. Lund, R. Crum, and M. Klagsbrun, Angiogenesis is stimulated by a tumor-derived endothelial cell growth factor, J Cell Biochem. 29 (1985) 275-87
181. R. Montesano, J.D. Vassalli, A. Baird, R. Guillemin, and L. Orci, Basic fibroblast growth factor induces angiogenesis in vitro, Proc Natl Acad Sci U S A. 83 (1986) 7297-301
182. M. Relf, S. LeJeune, P.A. Scott, S. Fox, K. Smith, R. Leek, A. Moghaddam, R. Whitehouse, R. Bicknell, and A.L. Harris, Expression of the angiogenic factors vascular endothelial cell growth factor, acidic and basic fibroblast growth factor, tumor growth factor beta-1, platelet-derived endothelial cell growth factor, placenta growth factor, and pleiotrophin in human primary breast cancer and its relation to angiogenesis, Cancer Res. 57 (1997) 963-9
183. A.L. Vescovi, B.A. Reynolds, D.D. Fraser, and S. Weiss, bFGF regulates the proliferative fate of unipotent (neuronal) and bipotent (neuronal/astroglial) EGF-generated CNS progenitor cells, Neuron. 11 (1993) 951-66
184. J.F. Reilly and V.G. Kumari, Alterations in fibroblast growth factor receptor expression following brain injury, Exp Neurol. 140 (1996) 139-50
185. I. Mocchetti, S.J. Rabin, A.M. Colangelo, S.R. Whittemore, and J.R. Wrathall, Increased basic fibroblast growth factor expression following contusive spinal cord injury, Exp Neurol. 141 (1996) 154-64
186. P. Follesa, J.R. Wrathall, and I. Mocchetti, Increased basic fibroblast growth factor mRNA following contusive spinal cord injury, Brain Res Mol Brain Res. 22 (1994) 1-8
187. K. Nozaki, S.P. Finklestein, and M.F. Beal, Basic fibroblast growth factor protects against hypoxia-ischemia and NMDA neurotoxicity in neonatal rats, J Cereb Blood Flow Metab. 13 (1993) 221-8
188. T.T. Lee, B.A. Green, W.D. Dietrich, and R.P. Yezierski, Neuroprotective effects of basic fibroblast growth factor following spinal cord contusion injury in the rat, J Neurotrauma. 16 (1999) 347-56
189. D.J. Martens, R.M. Seaberg, and D. van der Kooy, In vivo infusions of exogenous growth factors into the fourth ventricle of the adult mouse brain increase the proliferation of neural progenitors around the fourth ventricle and the central canal of the spinal cord, Eur J Neurosci. 16 (2002) 1045-57
190. K. Bendfeldt, V. Radojevic, J. Kapfhammer, and C. Nitsch, Basic fibroblast growth factor modulates density of blood vessels and preserves tight junctions in organotypic cortical cultures of mice: a new in vitro model of the blood-brain barrier, J Neurosci. 27 (2007) 3260-7
191. B. Reuss, R. Dono, and K. Unsicker, Functions of fibroblast growth factor (FGF)-2 and FGF-5 in astroglial differentiation and blood-brain barrier permeability: evidence from mouse mutants, J Neurosci. 23 (2003) 6404-12
127
192. W.R. Schabitz, F. Li, K. Irie, B.W. Sandage, Jr., K.W. Locke, and M. Fisher, Synergistic effects of a combination of low-dose basic fibroblast growth factor and citicoline after temporary experimental focal ischemia, Stroke. 30 (1999) 427-31; discussion 431-2
193. J. Bogousslavsky, S.J. Victor, E.O. Salinas, A. Pallay, G.A. Donnan, C. Fieschi, M. Kaste, J.M. Orgogozo, A. Chamorro, and A. Desmet, Fiblast (trafermin) in acute stroke: results of the European-Australian phase II/III safety and efficacy trial, Cerebrovasc Dis. 14 (2002) 239-51
194. G. Seghezzi, S. Patel, C.J. Ren, A. Gualandris, G. Pintucci, E.S. Robbins, R.L. Shapiro, A.C. Galloway, D.B. Rifkin, and P. Mignatti, Fibroblast growth factor-2 (FGF-2) induces vascular endothelial growth factor (VEGF) expression in the endothelial cells of forming capillaries: an autocrine mechanism contributing to angiogenesis, J Cell Biol. 141 (1998) 1659-73
195. R. Cao, E. Brakenhielm, R. Pawliuk, D. Wariaro, M.J. Post, E. Wahlberg, P. Leboulch, and Y. Cao, Angiogenic synergism, vascular stability and improvement of hind-limb ischemia by a combination of PDGF-BB and FGF-2, Nat Med. 9 (2003) 604-13
196. M. Fisher, M.E. Meadows, T. Do, J. Weise, V. Trubetskoy, M. Charette, and S.P. Finklestein, Delayed treatment with intravenous basic fibroblast growth factor reduces infarct size following permanent focal cerebral ischemia in rats, J Cereb Blood Flow Metab. 15 (1995) 953-9
197. S.E. Epstein, S. Fuchs, Y.F. Zhou, R. Baffour, and R. Kornowski, Therapeutic interventions for enhancing collateral development by administration of growth factors: basic principles, early results and potential hazards, Cardiovasc Res. 49 (2001) 532-42
198. J. Namiki, A. Kojima, and C.H. Tator, Effect of brain-derived neurotrophic factor, nerve growth factor, and neurotrophin-3 on functional recovery and regeneration after spinal cord injury in adult rats, J Neurotrauma. 17 (2000) 1219-31
199. L. Reynolds, R. Rauck, L. Webster, S. DuPen, E. Heinze, R. Portenoy, N. Katz, S. Charapata, M. Wallace, and D.M. Fisher, Relative analgesic potency of fentanyl and sufentanil during intermediate-term infusions in patients after long-term opioid treatment for chronic pain, Pain. 110 (2004) 182-8
200. T.L. Yaksh, ed. Spinal Drug Delivery. First ed. 1999, Elsevier: San Diego, CA.
201. ICU Equipment: Intravenous Line. 2002 [cited 2010 February 7]; Available from: http://www.waiting.com/intravenous.html.
202. N.Y. Times. Lumbar puncture. 2007 [cited 2010 February 7]; Available from: http://www.nytimes.com/imagepages/2007/08/01/health/adam/9587Lumbarpuncture.html.
203. T. Coradin, M. Boissiere, and J. Livage, Sol-gel chemistry in medicinal science, Curr Med Chem. 13 (2006) 99-108
128
204. R. Langer, L.G. Cima, J.A. Tamada, and E. Wintermantel, Future directions in biomaterials, Biomaterials. 11 (1990) 738-45
205. M.J. During, A. Freese, B.A. Sabel, W.M. Saltzman, A. Deutch, R.H. Roth, and R. Langer, Controlled release of dopamine from a polymeric brain implant: in vivo characterization, Ann Neurol. 25 (1989) 351-6
206. Y. Luo, K.R. Kirker, and G.D. Prestwich, Cross-linked hyaluronic acid hydrogel films: new biomaterials for drug delivery, J Control Release. 69 (2000) 169-84
207. J.R. Fried, Polymer Science and Technology, Second ed. Prentice Hall, Upper Saddle River, NJ, 2003
208. A.S. Hoffman, "Intelligent" polymers in medicine and biotechnology, Artif Organs. 19 (1995) 458-67
209. M.D. Baumann, C.E. Kang, J.C. Stanwick, Y. Wang, H. Kim, Y. Lapitsky, and M.S. Shoichet, An injectable drug delivery platform for sustained combination therapy, J Control Release. 138 (2009) 205-13
210. V.R. Sinha and A. Trehan, Biodegradable microspheres for protein delivery, J Control Release. 90 (2003) 261-80
211. A. Gutowska, B. Jeong, and M. Jasionowski, Injectable gels for tissue engineering, Anat Rec. 263 (2001) 342-9
212. N.A. Peppas, P. Bures, W. Leobandung, and H. Ichikawa, Hydrogels in pharmaceutical formulations, Eur J Pharm Biopharm. 50 (2000) 27-46
213. D.L. Elbert, A.B. Pratt, M.P. Lutolf, S. Halstenberg, and J.A. Hubbell, Protein delivery from materials formed by self-selective conjugate addition reactions, J Control Release. 76 (2001) 11-25
214. J.L. West and J.A. Hubbell, Comparison of covalently and physically cross-linked polyethylene glycol-based hydrogels for the prevention of postoperative adhesions in a rat model, Biomaterials. 16 (1995) 1153-6
215. Y. Katayama, Prolonged release of BDNF from PLGA Microspheres Dispersed within a PEG hydrogel, in Chemical Engineering. 2003, University of Toronto: Toronto.
216. C.E. Astete and C.M. Sabliov, Synthesis and characterization of PLGA nanoparticles, J Biomater Sci Polym Ed. 17 (2006) 247-89
217. P. Caliceti and F.M. Veronese, Pharmacokinetic and biodistribution properties of poly(ethylene glycol)-protein conjugates, Adv Drug Deliv Rev. 55 (2003) 1261-77
218. K.P. Vercruysse and G.D. Prestwich, Hyaluronate derivatives in drug delivery, Crit Rev Ther Drug Carrier Syst. 15 (1998) 513-55
129
219. E.A. Belazs, Medical Applications of Hyaluronan and its Derivatives. Cosmetic and Pharmaceutical Applications of Polymers, ed. C. Gebelein. Plenum Press, New York, Y, 1991. 293-310
220. K. Tomihata and Y. Ikada, Crosslinking of hyaluronic acid with water-soluble carbodiimide, J Biomed Mater Res. 37 (1997) 243-51
221. R.C. Savani, G. Cao, P.M. Pooler, A. Zaman, Z. Zhou, and H.M. DeLisser, Differential involvement of the hyaluronan (HA) receptors CD44 and receptor for HA-mediated motility in endothelial cell function and angiogenesis, J Biol Chem. 276 (2001) 36770-8
222. K.M. Sheehan, L.B. DeLott, S.M. Day, and D.H. DeHeer, Hyalgan has a dose-dependent differential effect on macrophage proliferation and cell death, J Orthop Res. 21 (2003) 744-51
223. L. Soltes, R. Mendichi, G. Kogan, J. Schiller, M. Stankovska, and J. Arnhold, Degradative action of reactive oxygen species on hyaluronan, Biomacromolecules. 7 (2006) 659-68
224. L. Li, P.M. Thangamathesvaran, C.Y. Yue, K.C. Tam, X. Hu, and Y.C. Lam, Gel network structure of methylcellulose in water, Langmuir. 17 (2001) 8062
225. Y. Xu, L. Li, P. Zheng, Y.C. Lam, and X. Hu, Controllable gelation of methylcellulose by a salt mixture, Langmuir. 20 (2004) 6134-8
226. Y. Xu, C. Wang, K.C. Tam, and L. Li, Salt-assisted and salt-suppressed sol-gel transitions of methylcellulose in water, Langmuir. 20 (2004) 646-52
227. A.K. Cheung and L.W. Henderson, Effects of complement activation by hemodialysis membranes, Am J Nephrol. 6 (1986) 81-91
228. M. Gardinali, M. Circardi, A. Agostoni, and T.E. Hugli, Complement activation in extracorporeal circulation: physiological and pathological implications, Pathol Immunopathol Res. 5 (1986) 352-70
229. F. Dumler and N.W. Levin, Membrane biocompatibility: clinical significance and therapeutic implications, Int J Artif Organs. 8 (1985) 257-62
230. Y. Rousseau, M.P. Carreno, J.L. Poignet, M.D. Kazatchkine, and N. Haeffner-Cavaillon, Dissociation between complement activation, integrin expression and neutropenia during hemodialysis, Biomaterials. 20 (1999) 1959-67
231. N.A. Hoenich, C. Woffindin, S. Stamp, S.J. Roberts, and J. Turnbull, Synthetically modified cellulose: an alternative to synthetic membranes for use in haemodialysis? Biomaterials. 18 (1997) 1299-303
232. M.C. Tate, D.A. Shear, S.W. Hoffman, D.G. Stein, and M.C. LaPlaca, Biocompatibility of methylcellulose-based constructs designed for intracerebral gelation following experimental traumatic brain injury, Biomaterials. 22 (2001) 1113-23
130
233. M.R. Wells, K. Kraus, D.K. Batter, D.G. Blunt, J. Weremowitz, S.E. Lynch, H.N. Antoniades, and H.A. Hansson, Gel matrix vehicles for growth factor application in nerve gap injuries repaired with tubes: a comparison of biomatrix, collagen, and methylcellulose, Exp Neurol. 146 (1997) 395-402
234. M.J. Roberts, M.D. Bentley, and J.M. Harris, Chemistry for peptide and protein PEGylation, Adv Drug Deliv Rev. 54 (2002) 459-76
235. C. Delgado, G.E. Francis, and D. Fisher, The uses and properties of PEG-linked proteins, Crit Rev Ther Drug Carrier Syst. 9 (1992) 249-304
236. A. Abuchowski, T. Vanes, N.C. Palczuk, R. Chen, and F.F. Davis, Effect of Covalently Attached Polyethylene-Glycol on Immunogenicity of Enzymes, Federation Proceedings. 36 (1977) 867-867
237. K.A. Witt, T.J. Gillespie, J.D. Huber, R.D. Egleton, and T.P. Davis, Peptide drug modifications to enhance bioavailability and blood-brain barrier permeability, Peptides. 22 (2001) 2329-43
238. J.M. Harris, N.E. Martin, and M. Modi, Pegylation: a novel process for modifying pharmacokinetics, Clin Pharmacokinet. 40 (2001) 539-51
239. C.E. Kang, C.H. Tator, and M.S. Shoichet, Conjugation of Poly(ethylene glycol) to Fibroblast Growth Factor 2 Enhances Distribution in Spinal Cord Tissue, Journal of Controlled Release. (2009) In Press
240. C.E. Kang, P.C. Poon, C.H. Tator, and M.S. Shoichet, A new paradigm for local and sustained release of therapeutic molecules to the injured spinal cord for neuroprotection and tissue repair, Tissue Eng Part A. 15 (2009) 595-604
241. C.E. Kang, R. Clarkson, C.H. Tator, W.T.I. Yeung, and M.S. Shoichet, Spinal Cord Blood Flow and Permeability Measured by Dynamic Computed Tomography Imaging in Live Animals After Localized Delivery of Fibroblast Growth Factor, Journal of Neurotrauma. (2010) In preparation
242. C.H. Tator, Review of treatment trials in human spinal cord injury: issues, difficulties, and recommendations, Neurosurgery. 59 (2006) 957-82; discussion 982-7
243. H. Terada, T. Kazui, M. Takinami, K. Yamashita, N. Washiyama, and B.A. Muhammad, Reduction of ischemic spinal cord injury by dextrorphan: comparison of several methods of administration, J Thorac Cardiovasc Surg. 122 (2001) 979-85
244. T.L. Yaksh, K.A. Horais, N. Tozier, M. Rathbun, P. Richter, S. Rossi, M. Grafe, C. Tong, C. Meschter, J.M. Cline, and J. Eisenach, Intrathecal ketorolac in dogs and rats, Toxicol Sci. 80 (2004) 322-34
245. S. Constantini and W. Young, The effects of methylprednisolone and the ganglioside GM1 on acute spinal cord injury in rats, J Neurosurg. 80 (1994) 97-111
131
246. S.K. Ray, D.D. Matzelle, E.A. Sribnick, M.K. Guyton, J.M. Wingrave, and N.L. Banik, Calpain inhibitor prevented apoptosis and maintained transcription of proteolipid protein and myelin basic protein genes in rat spinal cord injury, J Chem Neuroanat. 26 (2003) 119-24
247. K.A. Kigerl, W. Lai, S. Rivest, R.P. Hart, A.R. Satoskar, and P.G. Popovich, Toll-like receptor (TLR)-2 and TLR-4 regulate inflammation, gliosis, and myelin sparing after spinal cord injury, J Neurochem. 102 (2007) 37-50
248. M. Tamura, M. Nakamura, Y. Ogawa, Y. Toyama, M. Miura, and H. Okano, Targeted expression of anti-apoptotic protein p35 in oligodendrocytes reduces delayed demyelination and functional impairment after spinal cord injury, Glia. 51 (2005) 312-21
249. F. Qiao, C. Atkinson, H. Song, R. Pannu, I. Singh, and S. Tomlinson, Complement plays an important role in spinal cord injury and represents a therapeutic target for improving recovery following trauma, Am J Pathol. 169 (2006) 1039-47
250. R. Pannu and I. Singh, Pharmacological strategies for the regulation of inducible nitric oxide synthase: neurodegenerative versus neuroprotective mechanisms, Neurochem Int. 49 (2006) 170-82
251. G.D. Prestwich, D.M. Marecak, J.F. Marecek, K.P. Vercruysse, and M.R. Ziebell, Controlled chemical modification of hyaluronic acid: synthesis, applications, and biodegradation of hydrazide derivatives, J Control Release. 53 (1998) 93-103
252. G.T. Hermanson, Bioconjugate Techniques, Second ed. Academic Press, San Diego, CA, 1996
253. P.L. Ritger and N.A. Peppas, A simple equation for description of solute release I. Fickian and non-fickian release from non-swellable devices in the form of slabs, spheres, cylinders or discs, Journal of Controlled Release. 5 (1987) 23-36
254. B. Narasimhan and N.A. Peppas, Molecular analysis of drug delivery systems controlled by dissolution of the polymer carrier, Journal of Pharmaceutical Sciences. 86 (1997) 297-304
255. B.K. Kwon, W. Tetzlaff, J.N. Grauer, J. Beiner, and A.R. Vaccaro, Pathophysiology and pharmacologic treatment of acute spinal cord injury, Spine J. 4 (2004) 451-64
256. C. Mann, J.H. Lee, J. Liu, A.M. Stammers, H.M. Sohn, W. Tetzlaff, and B.K. Kwon, Delayed treatment of spinal cord injury with erythropoietin or darbepoetin--a lack of neuroprotective efficacy in a contusion model of cord injury, Exp Neurol. 211 (2008) 34-40
257. W.M. Deen, Analysis of Transport Phenomena. Topics in Chemical Engineering, ed. K.E. Gubbins. Oxford University Press, New York, 1998. 86-91
132
258. M. Stroh, W.R. Zipfel, R.M. Williams, W.W. Webb, and W.M. Saltzman, Diffusion of nerve growth factor in rat striatum as determined by multiphoton microscopy, Biophys J. 85 (2003) 581-8
259. B. Delpech, P. Bertrand, and C. Chauzy, An indirect enzymoimmunological assay for hyaluronidase, J Immunol Methods. 104 (1987) 223-9
260. U.B. Laurent, T.C. Laurent, L.K. Hellsing, L. Persson, M. Hartman, and K. Lilja, Hyaluronan in human cerebrospinal fluid, Acta Neurol Scand. 94 (1996) 194-206
261. F. Stohlman, Jr., S. Ebbe, B. Morse, D. Howard, and J. Donovan, Regulation of erythropoiesis. XX. Kinetics of red cell production, Ann N Y Acad Sci. 149 (1968) 156-72
262. M. Yalcin, F. Ak, I.T. Cangul, and M. Erturk, The effect of centrally administered erythropoietin on cardiovascular and respiratory system of anaesthetized rats, Auton Neurosci. 134 (2007) 1-7
263. G. Chen, J.X. Shi, C.H. Hang, W. Xie, J. Liu, and X. Liu, Inhibitory effect on cerebral inflammatory agents that accompany traumatic brain injury in a rat model: a potential neuroprotective mechanism of recombinant human erythropoietin (rhEPO), Neurosci Lett. 425 (2007) 177-82
264. I. Yatsiv, N. Grigoriadis, C. Simeonidou, P.F. Stahel, O.I. Schmidt, A.G. Alexandrovitch, J. Tsenter, and E. Shohami, Erythropoietin is neuroprotective, improves functional recovery, and reduces neuronal apoptosis and inflammation in a rodent model of experimental closed head injury, Faseb J. 19 (2005) 1701-3
265. C.H. Tator and M.G. Fehlings, Review of the secondary injury theory of acute spinal cord trauma with emphasis on vascular mechanisms, J Neurosurg. 75 (1991) 15-26
266. D.P. Ankeny, D.M. McTigue, Z. Guan, Q. Yan, O. Kinstler, B.T. Stokes, and L.B. Jakeman, Pegylated brain-derived neurotrophic factor shows improved distribution into the spinal cord and stimulates locomotor activity and morphological changes after injury, Exp Neurol. 170 (2001) 85-100
267. M. Stroh, W.R. Zipfel, R.M. Williams, S.C. Ma, W.W. Webb, and W.M. Saltzman, Multiphoton microscopy guides neurotrophin modification with poly(ethylene glycol) to enhance interstitial diffusion, Nat Mater. 3 (2004) 489-94
268. S.B. Zimmerman and L.D. Murphy, Electrophoresis of polyethylene glycols and related materials as sodium dodecyl sulfate complexes, Anal Biochem. 234 (1996) 190-3
269. A.N. Plotnikov, S.R. Hubbard, J. Schlessinger, and M. Mohammadi, Crystal structures of two FGF-FGFR complexes reveal the determinants of ligand-receptor specificity, Cell. 101 (2000) 413-24
133
270. Q. Yun, W.C. Xing, G.G. Mal, and Z.G. Su, Preparation and characterization of mono-PEGylated consensus interferon by a novel polyethylene glycol derivative, Journal of Chemical Technology and Biotechnology. 81 (2006) 776-781
271. M.F. Haller and W.M. Saltzman, Localized delivery of proteins in the brain: can transport be customized? Pharm Res. 15 (1998) 377-85
272. C. Nicholson, Diffusion and related transport mechanisms in brain tissue, Reports on Progress in Physics. 64 (2001) 815-884
273. M.J. Mahoney and W.M. Saltzman, Controlled release of proteins to tissue transplants for the treatment of neurodegenerative disorders, J Pharm Sci. 85 (1996) 1276-81
274. M. Mak, L. Fung, J.F. Strasser, and W.M. Saltzman, Distribution of drugs following controlled delivery to the brain interstitium, J Neurooncol. 26 (1995) 91-102
275. C.M. Bernards and H.F. Hill, Morphine and Alfentanil Permeability through the Spinal Dura, Arachnoid, and Pia Mater of Dogs and Monkeys, Anesthesiology. 73 (1990) 1214-1219
276. M.E. Young, P.A. Carroad, and R.L. Bell, Estimation of Diffusion-Coefficients of Proteins, Biotechnology and Bioengineering. 22 (1980) 947-955
277. A.N. Sandler and C.H. Tator, Review of the measurement of normal spinal cord blood flow, Brain Res. 118 (1976) 181-98
278. H. Toda, H. Maruyama, B. Budgell, and M. Kurosawa, Responses of dorsal spinal cord blood flow to noxious mechanical stimulation of the skin in anesthetized rats, J Physiol Sci. 58 (2008) 263-70
279. A.M. Peters, J. Brown, G.G. Hartnell, M.J. Myers, C. Haskell, and J.P. Lavender, Non-invasive measurement of renal blood flow with 99mTc DTPA: comparison with radiolabelled microspheres, Cardiovasc Res. 21 (1987) 830-4
280. A.M. Peters, R.D. Gunasekera, B.L. Henderson, J. Brown, J.P. Lavender, M. De Souza, J.M. Ash, and D.L. Gilday, Noninvasive measurement of blood flow and extraction fraction, Nucl Med Commun. 8 (1987) 823-37
281. R. Hebel and M.W. Stromberg, Anatomy and Embryology of the Laboratory Rat, 2nd ed. BioMed Verlag, Worthsee, 1986. 261
282. X.H. Wu, S.H. Yang, D.Y. Duan, H.H. Cheng, Y.T. Bao, and Y. Zhang, Anti-apoptotic effect of insulin in the control of cell death and neurologic deficit after acute spinal cord injury in rats, J Neurotrauma. 24 (2007) 1502-12
283. M. Murakami and M. Simons, Fibroblast growth factor regulation of neovascularization, Curr Opin Hematol. 15 (2008) 215-20
134
284. D.N. Loy, C.H. Crawford, J.B. Darnall, D.A. Burke, S.M. Onifer, and S.R. Whittemore, Temporal progression of angiogenesis and basal lamina deposition after contusive spinal cord injury in the adult rat, J Comp Neurol. 445 (2002) 308-24
285. P. Cuevas, F. Carceller, S. Ortega, M. Zazo, I. Nieto, and G. Gimenez-Gallego, Hypotensive activity of fibroblast growth factor, Science. 254 (1991) 1208-10
286. M. Ziche and L. Morbidelli, Nitric oxide and angiogenesis, J Neurooncol. 50 (2000) 139-48
287. D. Presti and J.E. Scott, Hyaluronan-mediated protective effect against cell damage caused by enzymatically produced hydroxyl (OH.) radicals is dependent on hyaluronan molecular mass, Cell Biochem Funct. 12 (1994) 281-8
288. V.C. Lees, T.P. Fan, and D.C. West, Angiogenesis in a delayed revascularization model is accelerated by angiogenic oligosaccharides of hyaluronan, Lab Invest. 73 (1995) 259-66
289. D.C. West and S. Kumar, The effect of hyaluronate and its oligosaccharides on endothelial cell proliferation and monolayer integrity, Exp Cell Res. 183 (1989) 179-96
290. T.S. Reese and M.J. Karnovsky, Fine structural localization of a blood-brain barrier to exogenous peroxidase, J Cell Biol. 34 (1967) 207-17
291. D.W. Beck, H.V. Vinters, M.N. Hart, and P.A. Cancilla, Glial cells influence polarity of the blood-brain barrier, J Neuropathol Exp Neurol. 43 (1984) 219-24
292. A. Yoshida, M. Yamamoto, T. Itoh, T. Irie, F. Hirayama, and K. Uekama, Utility of 2-Hydroxypropyl-Beta-Cyclodextrin in an Intramuscular Injectable Preparation of Nimodipine, Chemical & Pharmaceutical Bulletin. 38 (1990) 176-179
293. B.N. Nalluri, C. Milligan, J.H. Chen, P.A. Crooks, and A.L. Stinchcomb, In vitro release studies on matrix type transdermal drug delivery system of naltrexone and its acetyl prodrug, Drug Dev Ind Pharm. 31 (2005) 871-877
294. J. Siepmann, H. Kranz, R. Bodmeier, and N.A. Peppas, HPMC-matrices for controlled drug delivery: A new model combining diffusion, swelling, and dissolution mechanisms and predicting the release kinetics, Pharmaceut Res. 16 (1999) 1748-1756
295. H. Wen, K.R. Morris, and K. Park, Synergic effects of polymeric additives on dissolution and crystallization of acetaminophen, Pharmaceut Res. 25 (2008) 349-358
296. Z.G. He, D.F. Zhong, X.Y. Chen, X.H. Liu, X. Tang, and L.M. Zhao, Development of a dissolution medium for nimodipine tablets based on bioavailability evaluation, Eur J Pharm Sci. 21 (2004) 487-491
297. B.C. Hancock and M. Parks, What is the true solubility advantage for amorphous pharmaceuticals? Pharmaceut Res. 17 (2000) 397-404
135
298. M.E. Matteucci, B.K. Brettmann, T.L. Rogers, E.J. Elder, R.O. Williams, and K.P. Johnston, Design of potent amorphous drug nanoparticles for rapid generation of highly supersaturated media, Molecular Pharmaceutics. 4 (2007) 782-793
299. S.L. Raghavan, A. Trividic, A.F. Davis, and J. Hadgraft, Crystallization of hydrocortisone acetate: influence of polymers, Int J Pharm. 212 (2001) 213-221
300. Y. Wang, Y. Lapitsky, C.E. Kang, and M.S. Shoichet, Accelerated release of a sparingly soluble drug from an injectable hyaluronan-methylcellulose hydrogel, J Control Release. 140 (2009) 218-23
301. T.B. Csoka, G.I. Frost, and R. Stern, Hyaluronidases in tissue invasion, Invasion Metastasis. 17 (1997) 297-311
302. V. Deguine, M. Menasche, P. Ferrari, L. Fraisse, Y. Pouliquen, and L. Robert, Free radical depolymerization of hyaluronan by Maillard reaction products: role in liquefaction of aging vitreous, Int J Biol Macromol. 22 (1998) 17-22
303. H. Saari, Oxygen derived free radicals and synovial fluid hyaluronate, Ann Rheum Dis. 50 (1991) 389-92
304. C. Xu, E. Rosler, J. Jiang, J.S. Lebkowski, J.D. Gold, C. O'Sullivan, K. Delavan-Boorsma, M. Mok, A. Bronstein, and M.K. Carpenter, Basic fibroblast growth factor supports undifferentiated human embryonic stem cell growth without conditioned medium, Stem Cells. 23 (2005) 315-23
305. O.A. Ibrahimi, F. Zhang, S.C. Hrstka, M. Mohammadi, and R.J. Linhardt, Kinetic model for FGF, FGFR, and proteoglycan signal transduction complex assembly, Biochemistry. 43 (2004) 4724-30
306. J. Anderson, H.D. Burns, P. Enriquez-Harris, A.O. Wilkie, and J.K. Heath, Apert syndrome mutations in fibroblast growth factor receptor 2 exhibit increased affinity for FGF ligand, Hum Mol Genet. 7 (1998) 1475-83
307. W.Y. Chen and G. Abatangelo, Functions of hyaluronan in wound repair, Wound Repair Regen. 7 (1999) 79-89
308. H.G. Wisniewski, J.C. Hua, D.M. Poppers, D. Naime, J. Vilcek, and B.N. Cronstein, TNF/IL-1-inducible protein TSG-6 potentiates plasmin inhibition by inter-alpha-inhibitor and exerts a strong anti-inflammatory effect in vivo, J Immunol. 156 (1996) 1609-15
309. H. Kobayashi and T. Terao, Hyaluronic acid-specific regulation of cytokines by human uterine fibroblasts, Am J Physiol. 273 (1997) C1151-9
310. R. Cortivo, P. Brun, L. Cardarelli, M. O'Regan, M. Radice, and G. Abatangelo, Antioxidant effects of hyaluronan and its alpha-methyl-prednisolone derivative in chondrocyte and cartilage cultures, Semin Arthritis Rheum. 26 (1996) 492-501
136
311. P.W. Noble, Hyaluronan and its catabolic products in tissue injury and repair, Matrix Biol. 21 (2002) 25-9
312. A. Pinzon, A. Marcillo, D. Pabon, H.M. Bramlett, M.B. Bunge, and W.D. Dietrich, A re-assessment of erythropoietin as a neuroprotective agent following rat spinal cord compression or contusion injury, Exp Neurol. 213 (2008) 129-36
313. I. Koyanagi, C.H. Tator, and P.J. Lea, Three-dimensional analysis of the vascular system in the rat spinal cord with scanning electron microscopy of vascular corrosion casts. Part 1: Normal spinal cord, Neurosurgery. 33 (1993) 277-83; discussion 283-4
314. Y. Deguchi, T. Naito, T. Yuge, A. Furukawa, S. Yamada, W.M. Pardridge, and R. Kimura, Blood-brain barrier transport of 125I-labeled basic fibroblast growth factor, Pharm Res. 17 (2000) 63-9
315. B. Klein, W. Kuschinsky, H. Schrock, and F. Vetterlein, Interdependency of local capillary density, blood flow, and metabolism in rat brains, Am J Physiol. 251 (1986) H1333-40
316. Y. Hamamoto, T. Ogata, T. Morino, M. Hino, and H. Yamamoto, Real-time direct measurement of spinal cord blood flow at the site of compression: relationship between blood flow recovery and motor deficiency in spinal cord injury, Spine (Phila Pa 1976). 32 (2007) 1955-62
317. L.J. Noble and J.R. Wrathall, Blood-spinal cord barrier disruption proximal to a spinal cord transection in the rat: time course and pathways associated with protein leakage, Exp Neurol. 99 (1988) 567-78
318. L.J. Noble and J.R. Wrathall, The blood-spinal cord barrier after injury: pattern of vascular events proximal and distal to a transection in the rat, Brain Res. 424 (1987) 177-88
319. S.E. Sakiyama-Elbert and J.A. Hubbell, Development of fibrin derivatives for controlled release of heparin-binding growth factors, J Control Release. 65 (2000) 389-402
320. J.R. Tauro and R.A. Gemeinhart, Extracellular protease activation of chemotherapeutics from hydrogel matrices: a new paradigm for local chemotherapy, Mol Pharm. 2 (2005) 435-8
321. M.D. Baumann, C.E. Kang, C.H. Tator, and M.S. Shoichet, Safety and Biocompatibility of a Drug Delivery Platform for Sustained Combination Therapy to the Spinal Cord, Biomaterials. (2010) In preparation
322. A. Pinzon, A. Marcillo, A. Quintana, S. Stamler, M.B. Bunge, H.M. Bramlett, and W.D. Dietrich, A re-assessment of minocycline as a neuroprotective agent in a rat spinal cord contusion model, Brain Res. 1243 (2008) 146-51
137
Appendix 1: Glossary Angiogenesis: a physiological process involving the growth of new vessels from existing vessels.
Apoptosis: programmed cell death whereby cellular debris is limited and has less impact on
surrounding tissue. This is a genetically controlled process, although it can be affected by
biochemical events external to the cell.
Astrocytic scar: see Glial scar
Bioactivity: the ability of a molecule to perform its intended biological purpose.
Blood-spinal cord barrier: a cellular and molecular barrier that exists between blood and spinal
cord tissue; all blood vessels entering the CNS create this barrier which allows very little
molecular diffusion and passage from blood to tissue is limited by active transport mechanisms.
Cerebrospinal fluid: a clear fluid in the intrathecal space to cushion the brain and spinal cord.
Diffusivity: the extent to which a molecule can move through a given space.
Edema: abnormal fluid accumulation in the extracellular, extravascular space.
Elimination: removal of proteins from active diffusion, either through degradation or cellular
binding/uptake.
Endogenous: originating from within the organism of interest.
Equilibrium: A stable situation in which forces cancel one another; a state of dynamic systems in
which there is no net change.
Exogenous: originating from outside of the organism of interest.
Free Radical: molecules with unpaired electron shells which readily react and cleave other
molecules to fill their electron shells.
Glial Scar: a network of astrocytes that forms to seal healthy tissue from injury in the CNS.
Hypoxia: A deficiency of oxygen in biological tissues.
138
Intraperitoneal injection: injection into the body cavity (peritoneum). This is most commonly
performed in animals for systemic drug administration.
Intrathecal space: the space between the brain/spinal cord tissue and the membranes
(dura/arachnoid mater) that envelop the tissue; the space in which CSF flows.
Ischemia: A deficiency of oxygen due to a lack of blood flow to a tissue.
Immunogenicity: the ability of a molecule to induce an immune response in the body.
Necrosis: the premature death of cells caused by injury and infection; characterized by cellular
debris build-up in the surrounding area of dying cells.
Neuroprotection: mechanisms which protect neurons and neural tissue from degenerating after
injury or infection; characterized by sparing of neurons or axons and reduced lesion volumes.
Parenchyma: a general description of tissue including all cell types and tissue regions.
Penumbra: the area surrounding the traumatized tissue in which the secondary injury leads to
significant cell death and degeneration.
Secondary Injury: the cascade of biochemical events that occurs after mechanical trauma in
spinal cord tissue; characterized by inflammation, edema, hypoxia, ischemia, and cell death.
Shear thinning: a property of a material where viscosity temporarily decreases as shear rate
increases.
Systemic: affecting the entire bodily system, usually via the circulatory system.
Trophic: a growth-promoting property.
Vasoconstriction: narrowing of blood vessels due to contraction of muscular walls around the
vessels.
139
Appendix 2: Abbreviations
aCSF Artificial cerebrospinal fluid
ANOVA Analysis of variance
BBB Basso, Beattie, and Bresnahan open field locomotor scoring method
BSCB Blood-spinal cord barrier
CNS Central Nervous system
CSF Cerebrospinal fluid
DDS Drug delivery system
EDC 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride
ELISA Enzyme linked immunosorbent assay
EPO Erythropoietin
FGF2 Fibroblast growth factor 2 also known as basic Fibroblast growth factor
FITC Fluorescein isothiocyanate (fluorescent tag)
GFAP Glial fibrillary acidic protein
GM-1 Monosialotetrahexosylganglioside
GM-CSF Granulocyte macrophage colony stimulating factor
H&E Heamotoxylin and eosin
HA Hyaluronan
HAMC Hyaluronan and methylcellulose blend
140
IP Intraperitoneal
IT Intrathecal
IU International units
LFB Luxol fast blue
MC Methylcellulose
MP Methylprednisolone sodium succinate
MRI Magnetic resonance imaging
MW Molecular weight
NMR Nuclear magnetic resonance
NSCISC National Spinal Cord Injury Statistical Center
PBS Phosphate buffered saline
PEG Poly(ethylene glycol)
PEG-FGF2 Poly(ethylene glycol) conjugated to FGF2
PEG-mal Poly(ethylene glycol)-maleimide
PNS Peripheral Nervous system
PS Permeability – surface area product
SAS Subarachnoid space
SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis
SCBF Spinal cord blood flow
SCI Spinal cord injury
141
Sulfo-NHS N-hydroxsulfosuccinimide sodium salt
TM Tirilizad mesylate
T1/2 Thoracic vertebrae 1&2
TUNEL Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling
UV Ultraviolet
VEGF Vascular endothelial growth factor
142
Appendix 3: Additional Data
Figure 0.1: Protein release from HAMC is linearly related to time0.5
0
20
40
60
80
100
0 1 2 3 4 5 6 7 8 9
% R
elea
se
Time (h)
NBQXNim (EtOH)
0
20
40
60
80
100
0 1 2 3Time 0.5(h0.5)
for >60% release.
Figure 0.2: NBQX and solubilized nimodipine (with ethanol) release from HAMC. Inset
shows linear relationship with time0.5 for >60% of release.
143
Figure 0.3: Nimodipine release from HAMC for solid nimodipine particulates sonicated to
380nm and 900nm. Biphasic release profile was observed where fast release was observed
in the first 8 h (solid line = 380nm, dashed line = 900nm) and was likely due to soluble
nimodipine fraction saturating HAMC. Slow release was observed for 3 d (small dashed
line = 380 nm, large dashed line = 900 nm) likely due to fraction released as surface of
particles were solubilized into surrounding media.
144
Figure 0.4: Comparison of nimodipine solubility in MC and HA, with different nimodipine
particle sizes: (◆) 100-μm nimodipine particles in 0.25 wt% HA; and (□) 100-μm, (▲) 900-
nm, and (○) 380-nm nimodipine particles in 7 wt% MC. The upper shaded concentration
range indicates nimodipine solubility values achieved in 7 wt% MC, while the lower shaded
range indicates aqueous nimodipine solubility reported in the literature.
145
Figure 0.5: (A) Longitudinal echogram imaging of the spinal cord where brightness
indicates density of the tissue. The bright spots show the dense bones of the vertebrae
surrounding the spinal cord; yellow line indicates where the laminectomy was performed to
expose the spinal cord. The spinal cord tissue can be observed between the vertebrae only
below where the laminae have been removed. (B) Arterial flow from vessel in spinal cord.
(C) Venous flow from vessel in spinal cord.